GB2623743A

GB2623743A - A road cleaning vehicle

Info

Publication number: GB2623743A
Application number: GB2215294.6A
Authority: GB
Inventors: Knopp Gareth; Vajiramini Kuruppu Nanayakkara Ganindu; Offley Clive
Original assignee: Bucher Municipal Ltd
Current assignee: Bucher Municipal Ltd
Priority date: 2022-10-17
Filing date: 2022-10-17
Publication date: 2024-05-01
Also published as: GB202215294D0

Abstract

A road cleaning vehicle 10 comprises a driving system 16 and cleaning equipment 25. A control system comprises at least one debris sensor 60, 65 for generating first and second sensor data for locating debris 9 in front of and behind the vehicle respectively. A controller receives the first and second debris sensor data and processes the data in a debris identification model to identify debris. The driving system and/or the cleaning equipment are operated based upon the debris identified. The first sensor data may identify the location of debris and the second sensor data may identify debris erroneously not collected by the cleaning equipment, and the cleaning equipment or the steering of the vehicle may be adjusted accordingly. An alert system may provide information to an operator based on identified debris. The debris sensor may comprise a camera, lidar and/or a proximity sensor. Also claimed is a road cleaning vehicle having a navigation system and a control system for receiving debris map data indicative of at least one debris state associated with a predetermined location, the vehicle being operated according to the debris state.

Description

A ROAD CLEANING VEHICLE

Technical Field

The present disclosure is directed towards a road cleaning vehicle and a method of operating such a road cleaning vehicle.

Background

Road cleaning vehicles (also referred to as road sweepers, street cleaners and the like) are commonly used to remove unwanted debris from streets. A typical road cleaning vehicle is disclosed in GB2591486A and comprises road cleaning equipment such as cleaning brushes for sweeping the road, a debris collection system for drawing from the road into a hopper mounted on the vehicle chassis and water jets for cleaning the road. An operator in a cab controls the travel of the vehicle in addition to controlling the function of the road cleaning equipment.

There is an increasing interest in alternative powered Rechargeable Energy Storage System (RESS) vehicles, such as electrical vehicles (i.e. vehicles which do not use an internal combustion engine to provide a driving force), including road cleaning vehicles operating by electrical power rather than internal combustion engines. In electrical road cleaning vehicles, the range of the vehicle is limited by the energy storage capacity of the system, such as a battery, and the efficiency of its components. There is therefore now a need to increase the efficiency of operation of road cleaning vehicles, particularly in the control of the road cleaning equipment.

Furthermore, the need to control the travel of the road cleaning vehicle in conjunction with the road cleaning equipment places a high cognitive load on the operator, as well as requiring them to repetitively perform tasks. This may increase the fatigue of the operator and reduce the cleaning effectiveness of using the road cleaning vehicle. In addition, the operator needs a high level of skill and experience to be able to identify the type of debris on the road ahead and selectively operate the road cleaning equipment efficiently.

W02020/152526A1 discloses an automated road sweeper including a front facing video camera that scans the road ahead of the road sweeper to identify debris. The output from the video camera is used to select and exclusively operate the cleaning members covering a cleaning area in which the debris is located during movement of the sweeper. However, there is a continuing need to improve the efficiency and effectiveness of automated operation of all parts of the cleaning equipment.

Summary

An object of the present disclosure is to provide a road cleaning vehicle with improved efficiency and effectiveness of operation of the road cleaning equipment, such as by avoiding bulldozing over debris or unintentionally pushing such debris away from the road cleaning equipment and collection means.

The present invention therefore provides a road cleaning vehicle comprising: a driving system for driving the road cleaning vehicle along a road having debris thereon; road cleaning equipment for cleaning the debris on the road; and a control system comprising: at least one debris sensor configured for generating first debris sensor data for locating debris on the road in front of the road cleaning vehicle and second debris sensor data for locating debris on the road behind the road cleaning vehicle; and a controller configured to: receive first and second debris sensor data from the at least one debris sensor; process the first debris sensor data in a debris identification model to identify debris in front of the road cleaning vehicle and process the second debris sensor data in the debris identification model to identify any debris behind the road cleaning vehicle; and operate the driving system and/or road cleaning equipment based upon debris identified in front of and behind the road cleaning vehicle to control the cleaning of debris on the road by the road cleaning vehicle.

The present invention further provides a method of operating a road cleaning vehicle comprising, at the controller: receiving the first and second debris sensor data from the at least one debris sensor; processing the first debris sensor data in a debris identification model to identify debris in front of the road cleaning vehicle and process the second debris sensor data in the debris identification model to identify any debris behind the road cleaning vehicle; and operating the driving system and/or road cleaning equipment based upon the debris identified in front of and behind the road cleaning vehicle to control the cleaning of debris on the road by the road cleaning vehicle.

The present invention provides, in combination with or alternatively to the aforementioned vehicle, a road cleaning vehicle comprising: a driving system for driving the road cleaning vehicle along a road having debris thereon; road cleaning equipment for cleaning the debris on the road; a navigation system for determining a spatial location of the vehicle; and a control system comprising a controller configured to: determine the location and/or trajectory of the vehicle via the navigation system; receive debris map data indicative of at 3 -least one debris state associated with at least one predetermined location; determine from the location and/or trajectory and the debris map data that the vehicle is approaching and/or is at least partially in the at least one predetermined location; and operate the driving system and/or road cleaning equipment, as the vehicle travels at least partially over the at least one predetermined location, based upon the at least one debris state associated with the at least one predetermined location.

The present invention further provides, in combination with or alternatively to the aforementioned method, a method of operating the road cleaning vehicle comprising, at the controller: determining the location and/or trajectory of the vehicle via the navigation system; receiving debris map data indicative of at least one debris state associated with at least one predetermined location; determining from the location and/or trajectory and the debris map data that the vehicle is approaching and/or is at least partially in the at least one predetermined location; and operating the driving system and/or road cleaning equipment, as the vehicle travels at least partially over the at least one predetermined location, based upon the at least one debris state associated with the at least one predetermined location.

Brief Description of the Drawinas

By way of example only, embodiments of a road cleaning vehicle and a method of operating such a vehicle in accordance with the present disclosure are now described with reference to, and as shown in, the accompanying drawings, in which: FIGURE 1 is a side elevation view of a road cleaning vehicle in accordance with the present disclosure; FIGURE 2 is a schematic of a control system of the vehicle of Figure 1; FIGURE 3 is a schematic of a method of utilising data from first and second debris sensors of the vehicle of Figure 1; and FIGURE 4 is a top view of the vehicle of Figure 1 showing a particular use of the method of Figure 3.

Detailed Description

The present disclosure relates to a road cleaning machine including a plurality of sensors for automating the function of its road cleaning equipment. Front and rear facing sensors may be utilised to determine the effectiveness of the road cleaning equipment and adjust its used accordingly. 4 -

Road cleaning vehicles are commonly used to remove unwanted debris from streets. A road cleaning vehicle 10 in accordance with the present disclosure is shown in Figure 1. The road cleaning vehicle 10 in this instance is a two axled truck mounted sweeper 10 in the form of a driver operated vehicle and may be referred to as a road and/or street sweeper and/or cleaner. The vehicle 10 is for cleaning roads outdoors, preferably cleaning across a width of at least about 1000 mm, 2000 mm or 3000 mm in a single pass.

The vehicle 10 comprises an operator control station 20 for a human operator to use to control the road cleaning vehicle 10. The operator control station 20 may comprise at least one operator input 47, such as a steering wheel, buttons, levers, accelerator pedal, touchscreen or the like, for the operator to use to control the vehicle 10 (also known as a Human Machine Interface or HMI). The operator control station 20 may be located in an operator cab 13 in which a human operator can travel towards or at the front of the vehicle 10. The vehicle 10 may further comprise an alert system 48 for providing output and/or information to the operator. The alert system 48 may be located in the cab 13 and may form part of the operator control station 20. The alert system 48 may comprise a display, lights, dials or the like for displaying information to the operator. The alert system 48 may be controllable by the at least one operator input 47.

The vehicle 10 further comprises a chassis 19 to which the operator cab 13 and a debris container or hopper 17 may be mounted.

The vehicle 10 comprises a driving system 16 for driving the road cleaning vehicle 10 in a forward direction 7 along a road 18 having debris 9 thereon. The road cleaning vehicle 10 may also be operable travel in a rearward direction 21 and to perform road cleaning functions in the rearward direction 21. The terms "front" and "rear" in the present disclosure refer to the forwardmost and rearward most parts respectively of the vehicle 10 when travelling in forward direction 7.

The driving system 16 comprises a power system (not shown), which may be a RESS, and may comprise a Rechargeable Energy Storage System (RESS) and motor(s) and/or internal combustion engine and transmission. The power system is for driving the vehicle 10, such as by driving front and rear wheels 11, 12. The driving system 16 further comprises a steering system (not shown) for controlling the direction of the vehicle 10 along the road 18. The driving system 16 may be controlled by the operator via the operator control station 20.

-

The vehicle 10 further comprises road cleaning equipment 25 for cleaning the debris 9 on and/or from the road 18. The road cleaning equipment 25 may comprise at least one cleaning brush 14 for brushing the road 18, such as a pair of brushes 14, which may each be located on both outer sides of the vehicle 10 (only one is illustrated in Figure 1).

Alternatively, the road cleaning equipment 25 may comprise a plurality of brushes 14 on each side of the vehicle 10. The at least one cleaning brush 14 may be operable to perform brush functions such as rotating, raising from and lowering to the road 18 and/or tilting at differing angles relative to the road 18. The road cleaning equipment 25 may comprise a brush system 23 comprising the at least one cleaning brush 14 mounted to the chassis 19 by at least one brush mount 24. The at least one brush mount 24 may be configured to operate the at least one cleaning brush 14 to perform the brush functions, such as by comprising appropriate motors, actuators, servos and/or the like. The cleaning function of the brushes 14 is supported by the action of being able to apply pressure to the substrate with a selectable variable force.

The road cleaning equipment 25 may comprise a debris collection system 26 for picking up debris 9 from the road 18 and delivering it to the hopper 17. The debris collection system 26 comprises at least one inlet nozzle 15 and/or mechanical collection system, such as a vacuum airflow support mechanical elevator. The inlet nozzle 15 may in particular be a suction inlet nozzle 15 through which air and debris 9 is drawn.

The inlet nozzle 15 may be connected by at least one inlet conduit 27 to the hopper 17 and may comprise a inlet nozzle 15 and inlet conduit 27 on either side of the vehicle 10. The suction force in the conduits is provided by a fan of the debris collection system 26 that draws air and debris 9 through the at least one inlet nozzle 15 and is arranged to create a negative pressure in the hopper 17. The conveyancing force draws the debris 9 from the road 18, through the at least one inlet nozzle 15, through the at least one inlet conduit 27 and into the hopper 17. Once in the hopper 17, the debris 9 is separated from the air by means of a separation system before the air is exhausted by the centrifugal fan assembly to the atmosphere.

The at least one inlet nozzle 15 may be configured to perform inlet nozzle functions, such as being raised from and lowered to the road 18 and/or tilting at differing angles relative to the road 18. Figure 1 illustrates the at least one inlet nozzle 15 tilted backwardly so as to have a larger opening between the at least one inlet nozzle 15 and road 18 at the front of 6 -the at least one inlet nozzle 15 rather than its rear, thereby improving debris collection. The debris collection system 26 may comprises at least one inlet nozzle mount 28 mounting the least one inlet nozzle 15 to the chassis 19, vehicle 10 and/or hopper 17. The at least one inlet nozzle mount 28 may be configured to operate the at least one inlet nozzle 15 to perform the inlet nozzle functions, such as tilting and raising/lowering and may thus comprise appropriate motors, actuators, servos and/or the like.

The road cleaning equipment 25 may comprise a cleaning fluid system 30 for supplying cleaning fluid 31 to the road 18, to the vehicle 10 and/or within the vehicle 10, such as in a jet, spray and/or mist. The cleaning fluid 31 serves several functions, including as (a) a cleaning / wetting agent for the debris 9, (b) as lubrication within the debris collection system 26, such as to the at least one inlet conduit 27, to reduce both agglomeration of dirt and wear from erosion, and (c) as an atomised spray to attach to particulate matter (PM) and entrain it, thus adding mass to the particles which allows them to be separated from the airstream and be retained inside the hopper 17, thereby avoiding carryover through the fan and out into the atmospheric air.

The cleaning fluid 31 may be water, which may be stored in a tank mounted to the chassis 19. The cleaning fluid system 30 may be configured to supply the cleaning fluid 31 to the road 18 at or adjacent to the at least one brush 14, such as between the front wheel 11 and at least one brush 14, as illustrated. The cleaning fluid system 30 may additionally or alternatively be located at the front of the vehicle 10, i.e. in front of the front wheel 11, and/or at the rear of the vehicle 10, i.e. behind the rear wheel 12. The cleaning fluid system 30 may comprise at least one fluid supplier 32 (preferably a fluid supplier 32 on either side of the vehicle 10) directing cleaning fluid 31 at the road 18 as illustrated or, alternatively, may comprise an elongate bar extending across the width of the vehicle 10 for supplying a wide spray of cleaning fluid 31 across the vehicle 10. The fluid supplier 32 may be a nozzle, wand, bar or the like. The cleaning fluid system 30 may comprise a pump, valves and associated control systems (not shown) for controlling the supply of cleaning fluid 31.

The cleaning fluid system 30 may further comprise fluid suppliers or other supply means within the debris collection system 26, such as between the at least one inlet nozzle 15 and the hopper 17. Such internal fluid supplies are known as "PM10" jets and are commonly used to prevent carry over. 7 -

The vehicle 10 further comprises a control system 40, which is further illustrated in Figure 2. The control system 40 comprises a controller 41 for controlling the various parts of the vehicle 10. The controller 41 may be of any suitable known type and may comprise an engine control unit (ECU) or the like. The controller 41 may comprise multiple control units, such as an ECU, auxiliary ECU(s) or control unit(s), artificial intelligence control unit(s) possibly with effective parallel computing capabilities and the like. Generally, the controller 41 may comprise a memory 42 (e.g. multiple memories on different control units), which may store instructions or algorithms in the form of data, and a processing unit 43 (e.g. multiple processing units on different control units), which may be configured to perform operations based upon the instructions. The memory 42 may comprise one or more suitable computer-accessible or non-transitory storage mediums for storing computer program instructions, such as RAM, SDRAM, DDR SDRAM, RDRAM, SRAM, ROM, magnetic media, optical media and the like. The processing unit 43 may comprise one or more suitable processors capable of executing memory-stored instructions, such as a microprocessor, uniprocessor, a multiprocessor, a tensor processing unit (TPU) for machine learning and the like. The controller 41 may further comprise one or more graphics or tensor processing units (e.g. multiple graphics processing units on different control units) for rendering objects for viewing on a display and any other computing system equipment required for performing the functions of the vehicle 10.

The controller 41 may be communicatively connected (via a wired or wireless connection) to and/or exchange data with the at least one operator input 47, the operator control station 20, the alert system 48, the driving system 16, the road cleaning equipment 25, a navigation system 44, a communication system 45 and/or at least one sensor 60, 65. The controller 41 may receive data, such as from the at least one operator input 47, navigation system 44, communication system 45 and at least one sensor 60, 65, process the data to determine instructions and then perform operations based upon the instructions, such as by sending data to the alert system 48, driving system 16, road cleaning equipment 25 and/or communication system 45, performing calculations or carrying out logic-based tasks when doing so.

The control system 40 may comprise the communication system 45 for transferring data between the control system 40 and a remote computer system 50. The communication system 45 may comprise any type suitable apparatus for communication therebetween, particularly a wireless network. Exemplary wireless networks include a satellite communication network, broadband communication network, cellular, Bluetooth, 8 -microwave, point-to-point wireless, point-to-mulfipoint wireless, multipoint-to-mulfipoint wireless, Wireless Local Service (WIFi Dongle), Dedicated Short-Range Communications (DSRC) or any other wireless communication network. The remote computer system 50 may comprise a remote computer system memory 51 and remote computer system processing unit 52 and may, for example, be under the control of the owner or manufacturer of the vehicle 10.

The control system 40 may comprise the navigation system 44 for determining the position and/or pose (i.e. the current position of the vehicle 10 and its local pose relative to the local reference frame of the vehicle or the local map) of the vehicle 10. The navigation system 44 may determine the location of the vehicle 10 on the Earth's surface and/or may determine the location of the vehicle 10 relative to a reference position. The navigation system 44 may comprise any suitable navigation system 44, such as by determining the position of the vehicle 10 via a global navigation satellite system, such as global positioning system (GPS), or via triangulation with communication masts. The navigation system 44 may determine the location of the vehicle 10 by having an onboard map that may be interpreted with onboard sensors such as inertial navigation equipment to determine the current position of the vehicle 10 within the map. This could also be used to navigate the vehicle 10 to a different position within the map.

The controller 41 may be operable to control the driving system 16 to control the vehicle 10 speed and/or the direction of travel of the vehicle 10, such as by controlling the power system and/or steering system.

The controller 41 may be operable to control the road cleaning equipment 25 to control the cleaning of the road 18 by the vehicle 10. The controller 41 may be operable to control the brush system 23, such as the at least one cleaning brush 14 and/or at least one brush mount 24, to perform the brush functions. The controller 41 may be operable to control the debris collection system 26, such as the at least one inlet nozzle mount 28 and/or at least one inlet nozzle 15, to perform the inlet nozzle functions. The controller 41 may be operable to control the cleaning fluid system 30, such as the at least one fluid supplier 32 and pump, to control the supply of cleaning fluid 31 to the road 18.

The control system 40 may comprise a scanning system 59 for scanning the road 18 and for identifying debris 9 on the road 18. The scanning system 59 may comprise at least one first debris sensor 60 configured for generating first debris sensor data for locating debris 9 9 -on the road 18 in front of the vehicle 10 when moving in the forward direction 7 and/or at least one second debris sensor 65 configured for generating second debris sensor data for locating debris 9 on the road 18 behind the vehicle 10 when moving in the forward direction 7. The first and second debris sensors 60, 65 may be configured to generate the first and/or second debris sensor data relating to first and second scanning areas 61, 66 of the road 18 in front of and behind the vehicle 10 respectively. The first and second debris sensor data may be for identifying and/or locating debris 9 in the first and second scanning areas 61, 66, such as via processing of the sensor data in the controller 41.

In the present disclosure, the first and second debris sensors 60, 65 may be separate sensors or may be a single sensor (e.g. 360 degree camera), which may generate data in respect of an area covering both the first and second scanning areas 61, 66 together. Fundamentally, whether single or separate, the first and second debris sensors 60, 65 are capable of generating first and second debris sensor data in respect of the first and second scanning areas 61, 66 in front of and behind the vehicle 10.

The first and second scanning areas 61, 66 may be at least as wide as the vehicle 10 and may be up to twice or three times the width of the vehicle 10. The first and second scanning areas 61, 66 may extend along the road 18 away from the vehicle 10 by at least a quarter of the length of the vehicle 10 and optionally up to two times the length of the vehicle 10. The first and second scanning areas 61, 66 may be separated from the vehicle 10 (i.e. the first and second debris sensors 60, 65 do not generate sensor data relating to the road 18 adjacent to or at the front or rear of the vehicle 10). The first and second scanning areas 61,66 may be indicative of the effective cleaning area of the vehicle 10, (e.g. the possible swept space with reference to the path and location of the vehicle 10).

The effective cleaning area may be the local area within which the vehicle 10 can effect a change in the cleanliness level of the surrounding environment and scanned area. This would enable an effective measure of the cleaning performance of the vehicle 10 by distinguishing between what the vehicle 10 can and cannot change within these boundaries.

The first debris sensor 60 may be mounted at or towards the front of the vehicle 10 and may be directed downwards towards the road 18. The first debris sensor 60 may be mounted to the cab 13 and may be on the interior (such as by looking through the windscreen) or exterior (as illustrated) thereof. The second debris sensor 65 may be -10 -mounted at or towards the rear of the vehicle 10, such as to the hopper 17 and may be directed downwards towards the road 18.

The first and/or second debris sensors 60, 65, may each comprise at least one camera for capturing an image, such as of the scanning areas 61, 66, and the first and/or second debris sensor data may be image data. The first debris first and/or second debris sensor 60, 65, may comprise at least one of a camera, a 2D image camera, a 3D image camera, a monocular camera, an event camera, an infrared camera, a stereovision camera, lidar and/or a proximity sensor.

The first and/or second debris sensors 60, 65, may be configured to provide first and/or second debris data from which debris 9 can be identified in front of and/or behind the vehicle 10. In the present disclosure, the term "identify" in relation to debris 9 means the determination of the location of such debris 9 on the road 18 (particularly relative to the vehicle 10) and the type of (e.g. physical properties of) the debris 9 on the road 18.

The controller 41 may be configured to receive the first and second debris sensor data from the first and second debris sensors 60, 65 respectively. The controller 41 may be configured to process the first debris sensor data in a debris identification model to identify debris 9 in front of the vehicle 10 and process the second debris sensor data in the debris identification model to identify any debris 9 behind the vehicle 10. The controller 41 may be configured to operate the driving system 16 and/or road cleaning equipment 25 based upon debris 9 identified in front of and behind the vehicle 10 to control the cleaning of debris 9 on the road 18 by the vehicle 10.

By using both the first and second debris sensor data to operate the vehicle 10, the effectiveness of cleaning can be monitored and improved whilst the vehicle 10 is operating. In particular, the first debris sensor data can be used initially to control the driving system 16 and/or road cleaning equipment 25 to collect or clean the debris 9 identified in front of the vehicle 10. The second debris sensor data can be used as feedback to determine the effectiveness of such control by detecting if any debris 9 has been left behind the vehicle 10 and the control of the driving system 16 and/or road cleaning equipment 25 may be adjusted accordingly.

In particular, the controller 41, such as by the processing unit 43, may be configured to process the first and/or second debris sensor data in a debris identification model for identifying the debris 9 on the road 18 in front of and behind the vehicle 10. The debris identification model may be stored on the memory 42. The debris identification model may be updated based upon the first and/or second debris data and/or from data communicated to the control system 40 from the remote computer system 50 via the communication system 45.

The debris identification model may comprise any suitable algorithm for identifying and/or classifying debris. The debris identification model may comprise a neural network and may receive the first and/or second debris sensor data (e.g. one or more images of the first and second scanning areas 61, 66), detects objects, including debris 9, and their location from such data and then allocate the objects to one of a plurality of object classifications, including a collect classification indicating that there is debris 9 to be collected. One or more object classification(s) may be collect classification(s) if the object is debris 9, for indicating that the vehicle 10 should clean or collect the debris 9. One or more object classifications may be avoid classification(s), for example if the object is a human, for indicating that the vehicle 10 should avoid the object. There may be a plurality of collect classifications, with each collect classification indicating a certain type of debris 9, such as soft debris 9 (e.g. wrappers and foils) and hard debris 9 (e.g. bricks and stones). One or more object classification(s) may be ignore classification(s) for indicating that the vehicle 10 does not need to take any action (e.g. the surface of the road 18 may be classified as ignore as the vehicle 10 does not need to collect or avoid it).

Figure 3 illustrates a method 79 in accordance with the present disclosure comprising steps the controller 41 is configured to carry out. As the vehicle 10 drives along the road 18 (step 80), the first debris sensor 60 captures first debris sensor data including information relating to any debris 9 on the road 18 in the first scanning area 61 (step 81), which may correspond to a road area 78. The controller 41 receives the first debris sensor data and, via the processing unit 43, processes the first debris sensor data (step 82) using the debris identification model (step 83) to identify the debris 9 in the first scanning area 61. Such identification may include the type and/or location of the debris 9, preferably including which classification the debris 9 falls into (e.g. collect, avoid, ignore). However, such identification may simply be an identification that there is some type of debris 9 present and there may be a control action-or behaviour recommendation associated with such an identification.

-12 -The controller 41 may be configured to operate the driving system 16 and/or road cleaning equipment 25 based upon the debris 9 identified in front of the vehicle 10 using a vehicle cleaning control strategy. In particular, the vehicle cleaning control strategy may link the debris 9 identified in front of the vehicle 10 to operational settings of the driving system 16 and/or road cleaning equipment 25. The vehicle cleaning control strategy may be stored on the memory 42. The vehicle cleaning control strategy may be updated based upon the first and/or second debris data and/or from data communicated to the control system 40 from the remote computer system 50 via the communication system 45. The vehicle cleaning control strategy may also link environmental objects (i.e. not debris) around the vehicle 10, such as kerbs, other vehicles, road furniture or the like, which may for example be determined from the first and/or second debris sensor data (e.g. from the same camera image), to operational settings of the driving system 16 and/or road cleaning equipment 25.

As illustrated in Figure 3, the method 79 may comprise, and the controller 41 may be configured to, process the identified debris 9 (including its type and/or location) (step 84) in the vehicle cleaning control strategy (step 85) to determine the operational settings of the driving system 16 and/or road cleaning equipment 25. The controller 41 may further receive or determine trajectory data indicative of the vehicle 10 trajectory and use such trajectory data during such processing. Subsequently, the controller 41 may control the driving system 16 and/or road cleaning equipment 25 in accordance with the determined operational settings (step 86).

Prior to the control according to the determined operational settings (step 86), an automated functional safety supervision system 94 may perform diagnostics 95 to ensure that they comply with safety parameters, which may be predetermined. If the safety parameters are met or breached, alerts 96 may be generated for the operator such that the operator can take manual control. Alternatively, the vehicle cleaning control strategy may implement an alternative set of operational settings at step 84 until the operation is within the safety parameters.

For example, if the debris 9 is in the avoid classification, the controller 41 may adjust the settings (e.g. machine speed and travel direction) of the driving system 16 so the vehicle 10 drives around the debris 9. If the debris 9 is in the ignore classification, the controller 41 may stop the operation of the road cleaning equipment 25. If the debris 9 is in the collect classification, the controller 41 may control the road cleaning equipment 25 to clean and/or collect the debris 9. In particular, the controller 41 may selectively control the brush system -13 - 23, debris collection system 26 and/or cleaning fluid system 30 to clean and/or collect the debris 9 based upon the type and/or location of the identified debris 9. For example, if the brush 14 on the side of the debris 9 of the vehicle 10 may be activated or actuated to clean it. If the debris 9 is PM, the cleaning fluid system 30 may be activated or actuated to wash it. If the debris 9 is suitable for collection, the inlet nozzle 15 on the side of the vehicle 10 of the debris 9 may be activated or actuated to collect it in the hopper 17.

Subsequently, the vehicle 10 continues along the road 18 until the second debris sensor 65 and second scanning area 66 at least partially cover the same road area 78 as initially covered by the first scanning area 66 in step 81. It will be appreciated that such scanning continues repeatedly such that this scanning is repeated continuously, building up a continuous map of debris 9 along the road 18. The method 29 then comprises, at step 87, operating the second debris sensor 65 to capture second debris sensor data indicative of any debris 9 remaining in the second scanning area 66 and/or road area 78.

The second debris sensor data is then processed by the controller 41 (step 88) in the debris identification model (step 83) to identify any debris 9 remaining on the road 18 after the vehicle 10 has passed over the road area 78. The controller 41 may further receive or determine trajectory data indicative of the vehicle 10 trajectory and use such trajectory data during such processing. This provides feedback to enable the improved cleaning by the vehicle 10.

In particular, the controller 41 may be configured to operate the driving system 16 and/or road cleaning equipment 25 based upon the debris 9 identified behind the road cleaning vehicle 10. The controller 41 may adjust the operational settings of the driving system 16 and/or road cleaning equipment 25 (step 89) by using the vehicle cleaning control strategy (step 85) as the vehicle cleaning control strategy may also link debris 9 identified behind the vehicle 10 to the operational settings of the driving system 16 and/or road cleaning equipment 25 Figure 4 illustrates a particular embodiment of the present disclosure in which the road cleaning equipment 25 comprises a pair of brushes 14 for cleaning the road 18. In such circumstances, the vehicle 10 may pass over a wide area of debris 9a and the brushes 14 may not completely clean the debris 9a, leaving a strip 9b of uncleaned debris 9b behind the vehicle 10, such as due to the brushes 14 being too far apart. The controller 41 may be configured to, from the debris 9b identified in the second debris sensor 65 data, detect the -14 -strip 9b of uncleaned debris along the road 18 resulting from the gap between the brushes 14 from the debris 9b identified behind the vehicle 10. The controller 41 may be configured to, based upon the vehicle cleaning control strategy (step 85), operate the brushes 14 to close the gap between them (step 89) so as to prevent the vehicle 10 from leaving the strip 9b of uncleaned debris.

The controller 41 may also update the vehicle cleaning control strategy and/or debris identification model (step 90) based upon the debris 9 identified behind the vehicle 10. The vehicle cleaning control strategy and/or debris identification model may be updated based upon a comparison of the debris 9 identified in front of and behind the vehicle 10 and/or the operational settings of the cleaning equipment 25 at the time.

In particular, the debris 9 identified behind the vehicle 10 may be used in machine learning to update the vehicle cleaning control strategy and/or debris identification model, such as by operating as new training data for the vehicle cleaning control strategy and/or debris identification model. As a result, in future cleaning operations the road cleaning equipment 25 may be operated in a more efficient manner based upon the learning by the model(s) of the debris 9 left behind the vehicle 10 when the road cleaning equipment 25 was operated in certain circumstances.

The vehicle cleaning control strategy may comprise a machine learning algorithm to update the operational settings of the driving system 16 and/or road cleaning equipment 25 to optimise the future collection of debris 9. The machine learning algorithm may be model-based or model-free and may be a reinforcement learning algorithm. For example, the machine learning algorithm may be a 0-learning algorithm where there are weights and loss functions, which can tune the vehicle cleaning control strategy between exploration (i.e. trying new operational settings) and exploitation (using operational settings from already learnt optimisation). For example, the debris 9 identified behind the vehicle 10 may be classified according to reward levels and those reward levels would be used in the reinforcement learning algorithm to update the operational settings of the driving system 16 and/or road cleaning equipment 25 to optimise the future collection of debris 9.

The method 79 may further comprise the controller 41 generating an alert (step 91) via the alert system 48 to provide information to the operator indicative of the debris 9 left behind the vehicle 10. For example, the alert may be on a display and indicate to the operator that there is still debris 9 being left by the cleaning equipment 25. The controller 41 may -15 -determine an alert status based upon the debris 9 identified in front of and/or behind the vehicle 10 and operate the alert system 48 to generate an alert indicative of the alert status for the operator.

The method 79 may further comprise, as in steps 92, 93, the controller 41 monitoring an adjustment of the operational settings of the driving system 16 and/or road cleaning equipment 25 by the operator and subsequently updating the vehicle cleaning control strategy and/or debris identification model based upon the adjustment of the operational settings by the operator. In particular, the debris 9 identified behind the vehicle 10 and adjustment of the operational settings may be used in machine learning to update the vehicle cleaning control strategy and/or debris identification model, such as by operating as new training data for the vehicle cleaning control strategy and/or debris identification model. The vehicle cleaning control strategy may comprise an expert system or Q-learning system using, for example, Convolutional Neural Networks (CNNs) to gain insight from the images and update the strategy for controlling the operational settings accordingly. As a result, in future cleaning operations the road cleaning equipment 25 may be operated in a more efficient manner based upon the learning by the model(s) of how the operator dealt with the debris 9 left behind the vehicle 10 when the road cleaning equipment 25 was operated in certain circumstances.

The vehicle cleaning control strategy may further be adapted based upon the location of the vehicle 10 and debris map data. The debris map data may be representative of debris around a region, such as at least part of an urban area such as a town or city. The debris map data may indicate at least one debris state at or associated with at least one predetermined location. The debris state may comprise the type, amount and/or other properties of debris that is either known to be or has historically been found at the predetermined location. The debris state may be one of a range of values or be part of an index, such as between 0 and 1, with 1 indicating a state of more debris.

Such debris map data may be obtained via vehicles, whether road cleaning vehicles 10 or other types of vehicles, with debris sensors, such as cameras, travelling around the region to assess the debris state and thus generate and store the debris map data. The historical data can be obtained by sequentially obtaining such debris map data in the predetermined locations or areas over a period of time (e.g. at least a week of with a plurality of data collections). For example, by collecting debris map data and identifying debris states a plurality of times over multiple days, weeks and/or months, the debris map data may be -16 -indicative of debris collecting on certain streets at certain times (e.g. outside entertainment venues on certain, repetitive mornings of the week). Hence the debris map data may be indicative of predetermined locations or areas that require more or less cleaning by a road cleaning vehicle 10.

The debris map data may be stored in the memory 51 of the remote computer system 50 and communicated to the controller 41 of the vehicle 10. The controller 41 may determine the trajectory and/or location of the vehicle 10, such as via the navigation system 44, so as to spatially locate the vehicle 10 relative to the predetermined locations or areas.

The controller 10 may determine the debris state of an upcoming predetermined location based upon the debris may data and location and/or trajectory of the vehicle 10. The controller 41 may then determine the operational settings of the driving system 16 and/or road cleaning equipment 25 based upon the vehicle cleaning control strategy and the debris state of the upcoming predetermined location. This may be to improve cleaning of the predetermined location. For example, if the debris state indicates that the upcoming predetermined location historically has a significant amount of debris (e.g. a debris state index is close to 1), the road cleaning equipment 25 may be operated at a higher cleaning level than it would have been otherwise. Such control may be implemented by adjusting weightings applied to individual control of parts of the road cleaning equipment 25 in the vehicle cleaning control strategy in accordance with the debris state.

It will be apparent that such a control strategy does not necessarily require the scanning system 59. However, the scanning system 59 may be utilised in the generation of the debris map data by capturing images of the road 18 and debris 9 thereon in use and returning the image data to the remote computer system 50 to build up the debris map data. In addition, preferably the scanning system 59 is utilised with the debris map data to improve control of the vehicle 10 to optimise cleaning.

As a predetermined location enters the first scanning area 61, the controller 41 may compare the captured first debris data with the associated debris map data. In particular, the first debris data may be compared with the debris state so as to assess whether the debris state is accurate. If accurate, the controller 41 may then determine the operational settings based upon the vehicle cleaning control strategy and the debris state. If not accurate, the controller 41 may determine the operational settings based upon the vehicle cleaning control strategy and the first debris data without the debris state. Alternatively, if -17 -not accurate, the controller 41 may determine the operational settings based upon the vehicle cleaning control strategy, the first debris data and the debris state, with lower weighting being applied to the debris state than the first debris data when selecting the operational settings. If not accurate, the controller 41 may inform the remote computer system 50, which may update the debris map data on its memory 51.

The effectiveness of cleaning, determined by comparing the first and second debris data, may also be stored for future reference in the debris map data. For example, if it is determined that in a predetermined location the cleaning effectiveness was not sufficient, such as due to more debris than expected being left behind the vehicle 10, the debris state for that predetermined location may be marked and/or changed so as to indicate a higher amount of debris. Such effectiveness of cleaning may then be taken into account by the controller 41 when cleaning the predetermined location in future. In particular, the controller 41 may subsequently determine the operational settings for the predetermined location based upon the vehicle cleaning control strategy and the debris state, including the historical cleaning effectiveness at the predetermined location.

By using the debris map data, operation of the vehicle 10 can be improved by using historical data indicative of where there is a lot of debris to control the vehicle 10 appropriately to ensure that such debris is picked up. Furthermore, the vehicle 10 can capture suitable information to keep the debris map data up-to-date. In addition, if certain areas are particularly hard to clean, this can be captured from the second debris data and then higher cleaning settings can be utilised in future in those areas, thereby improving cleaning.

Claims

-18 -CLAIMS: 1 A road cleaning vehicle comprising: a driving system for driving the road cleaning vehicle along a road having debris thereon; road cleaning equipment for cleaning the debris on the road and a control system comprising: at least one debris sensor configured for generating first debris sensor data for locating debris on the road in front of the road cleaning vehicle and second debris sensor data for locating debris on the road behind the road cleaning vehicle; and a controller configured to: receive first and second debris sensor data from the at least one debris sensor; process the first debris sensor data in a debris identification model to identify debris in front of the road cleaning vehicle and process the second debris sensor data in the debris identification model to identify any debris behind the road cleaning vehicle; and operate the driving system and/or road cleaning equipment based upon debris identified in front of and behind the road cleaning vehicle to control the cleaning of debris on the road by the road cleaning vehicle.
2. A road cleaning vehicle as claimed in claim 1 wherein the controller is configured to: process the first debris sensor data in a debris identification model to identify debris in front of the road cleaning vehicle, such as the location of debris; operate the driving system and/or road cleaning equipment based upon debris identified in front of the road cleaning vehicle to control the cleaning of debris on the road by the road cleaning vehicle, such as operating the road cleaning equipment and/or steering the vehicle to collect and/or clean the debris; process the second debris sensor data in the debris identification model to identify debris behind of the road cleaning vehicle, such as debris erroneously not collected by the road cleaning equipment; and operate the driving system and/or road cleaning equipment based upon the debris identified behind the road cleaning vehicle to control the cleaning of debris on the road by the road cleaning vehicle, such as by adjusting the road cleaning equipment and/or steering the vehicle to avoid further debris erroneously not being collected and/or by adjusting a vehicle control strategy and/or debris identification model so as to avoid further debris erroneously not being collected during future cleaning.-19 -
3. A road cleaning vehicle as claimed in claim 1 or claim 2 wherein the at least one debris sensor comprises a first debris sensor configured for generating first debris sensor data for locating debris on the road in front of the road cleaning vehicle and a second debris sensor configured for generating second debris sensor data for locating debris on the road behind the road cleaning vehicle
4. A road cleaning vehicle as claimed in any one of the preceding claims wherein the controller is configured to operate the driving system and/or road cleaning equipment based upon the debris identified in front of the road cleaning vehicle using a vehicle cleaning control strategy.
5. A road cleaning vehicle as claimed in claim 3 wherein the controller is configured to operate the driving system and/or road cleaning equipment, using the vehicle cleaning control strategy, also based upon the location of the vehicle and debris map data, environmental objects around the vehicle and/or a determined trajectory of the vehicle.
6. A road cleaning vehicle as claimed in any one of the preceding claims wherein the controller is configured to operate the driving system and/or road cleaning equipment based upon the debris identified behind the road cleaning vehicle by: using the vehicle cleaning control strategy; and/or updating the vehicle cleaning control strategy and/or debris identification model based upon the debris identified behind the road cleaning vehicle.
7. A road cleaning vehicle as claimed in claim 6 wherein the vehicle cleaning control strategy and/or debris identification model is updated based upon a comparison of the debris identified in front of and behind the road cleaning vehicle.
8. A road cleaning vehicle as claimed in any one of the preceding claims wherein the road cleaning vehicle further comprises an alert system for providing information to an operator and the controller is further configured to: determine an alert status based upon the debris identified in front of and behind the road cleaning vehicle; and operate the alert system to generate an alert indicative of the alert status for the 35 operator.-20 -
9. A road cleaning vehicle as claimed in any one of the preceding claims wherein the controller is configured to monitor an adjustment of the operational settings of the driving system and/or road cleaning equipment by the operator and subsequently update the vehicle cleaning control strategy and/or debris identification model based upon the adjustment of the operational settings by the operator.
10. A road cleaning vehicle as claimed in any one of the preceding claims wherein the at least one debris sensor comprises at least one of a camera, a 2D image camera, a 3D image camera, a monocular camera, an event camera, an infrared camera, a stereovision camera, lidar and/or a proximity sensor.
11. A road cleaning vehicle as claimed in any one of claims 3 to 9 wherein the first debris sensor is configured to capture sensor data for locating debris in a first scanning area of the road in front of the vehicle and the second debris sensor is configured to capture sensor data for locating debris in a second scanning area of the road behind the road cleaning vehicle.
12. A road cleaning vehicle as claimed in any one of claims 3 to 11 wherein the road cleaning vehicle comprises a cab and the first debris sensor is mounted to, adjacent to or in the cab and/or the road cleaning vehicle comprises a hopper and the second debris sensor is mounted to, adjacent to or in the hopper.
13. A road cleaning vehicle as claimed in any one of the preceding claims wherein the road cleaning equipment comprises at least a pair of brushes for cleaning the road and the controller is configured to: process the second debris sensor data in the debris identification model to identify a line of uncleaned debris along the road behind the vehicle resulting from a gap between the brushes from the debris identified behind the road cleaning vehicle; and operate the driving system and/or road cleaning equipment based upon debris identified in front of and behind the road cleaning vehicle to control the cleaning of debris on the road by the road cleaning vehicle by operating at least the pair of brushes to close the gap between them so as to prevent the road cleaning vehicle from leaving the line of uncleaned debris.
14. A road cleaning vehicle as claimed in any one of the preceding claims wherein the at least one debris sensor comprises at least one of a camera, a 2D image camera, a 3D -21 -image camera, a monocular camera, an event camera, an infrared camera, a stereovision camera, lidar and/or a proximity sensor.
15. A method of operating a road cleaning vehicle of any one of the preceding claims comprising, at the controller: receiving the first and second debris sensor data from the at least one debris sensor; processing the first debris sensor data in a debris identification model to identify debris in front of the road cleaning vehicle and process the second debris sensor data in the debris identification model to identify any debris behind the road cleaning vehicle; and operating the driving system and/or road cleaning equipment based upon the debris identified in front of and behind the road cleaning vehicle to control the cleaning of debris on the road by the road cleaning vehicle.
16. A road cleaning vehicle comprising: a driving system for driving the road cleaning vehicle along a road having debris thereon; road cleaning equipment for cleaning the debris on the road, a navigation system for determining a spatial location of the vehicle; and a control system comprising a controller configured to: determine the location and/or trajectory of the vehicle via the navigation system; receive debris map data indicative of at least one debris state associated with at least one predetermined location; determine from the location and/or trajectory and the debris map data that the vehicle is approaching and/or is at least partially in the at least one predetermined location; and operate the driving system and/or road cleaning equipment, as the vehicle travels at least partially over the at least one predetermined location, based upon the at least one debris state associated with the at least one predetermined location.
17. A road cleaning vehicle according to claim 16 wherein the debris map data is representative of debris around a region and/or the at least one debris state comprises the type, amount and/or other properties of debris that is either known to be or has historically been found at the at least one predetermined location.-22 -
18. A road cleaning vehicle according to claim 16 or claim 17 further comprising at least one debris sensor for generating debris sensor data for locating debris on the road around the road cleaning vehicle, the debris data subsequently being utilised in updating the at least one debris state associated with at least one predetermined location in the debris map data.
19. A road cleaning vehicle according to any one of claims 16 to 18 further comprising at least one first debris sensor configured for generating first debris sensor data for locating debris on the road in front of the road cleaning vehicle, wherein the controller is configured to: compare the first debris data with the at least one debris state so as to assess whether the at least one debris state is accurate for at least one predetermined location; and based upon such comparison, determine whether to operate the driving system and/or road cleaning equipment, as the vehicle travels at least partially over the at least one predetermined location, based upon first debris data and/or the at least one debris state associated with the at least one predetermined location.
20. A road cleaning vehicle according to any one of claims 16 to 19 wherein the debris map data includes an indication of historical cleaning effectiveness at the predetermined location and the controller is configured to operate the driving system and/or road cleaning equipment, as the vehicle travels at least partially over the at least one predetermined location, based upon the historical cleaning effectiveness to control the cleaning of debris on the road by the road cleaning vehicle.
21. A method of operating the road cleaning vehicle of any one of claims 16 to 20 comprising, at the controller: determining the location and/or trajectory of the vehicle via the navigation system; receiving debris map data indicative of at least one debris state associated with at least one predetermined location; determining from the location and/or trajectory and the debris map data that the vehicle is approaching and/or is at least partially in the at least one predetermined location; and operating the driving system and/or road cleaning equipment, as the vehicle travels at least partially over the at least one predetermined location, based upon the at least one debris state associated with the at least one predetermined location.