CN110906946A - Service robot navigation planning method for distinguishing personnel influence - Google Patents
Service robot navigation planning method for distinguishing personnel influence Download PDFInfo
- Publication number
- CN110906946A CN110906946A CN201911211423.7A CN201911211423A CN110906946A CN 110906946 A CN110906946 A CN 110906946A CN 201911211423 A CN201911211423 A CN 201911211423A CN 110906946 A CN110906946 A CN 110906946A
- Authority
- CN
- China
- Prior art keywords
- cost
- personnel
- service robot
- influence
- user
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
- G01C21/34—Route searching; Route guidance
- G01C21/3446—Details of route searching algorithms, e.g. Dijkstra, A*, arc-flags, using precalculated routes
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Automation & Control Theory (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
The invention discloses a service robot navigation planning method for distinguishing personnel influence. The movement cost is obtained by geometric modeling using a gaussian function; the personnel influence model is brought into the environment map cost model, and is closely linked with the combined navigation planning process; and (4) checking whether the personnel participate according to the laser radar sensor, and distinguishing the roles of the personnel. If the detected user is the user, switching the thread, and navigating the plan to be close to the user; and if the pedestrian is detected, performing obstacle avoidance planning without invading the private space of the pedestrian. In addition, the combined planning process is divided into global and local parts, and the planning algorithm is optimized to be matched with the personnel influence cost. The invention relieves the processing dilemma of the existing service robot navigation planning technology when the coping personnel participate, and closely links different cost influences with the planning process, so that the robot really ' realizes ' the existence of people ' and greatly improves the user experience and the execution capacity of the service robot.
Description
Technical Field
The invention belongs to the technical field of mobile service robots, and particularly relates to a service robot navigation planning method for distinguishing personnel participation influence.
Background
Robots are increasingly entering our homes and workplaces, especially autonomous mobile robots that serve the role of assistance and service, greatly improving the quality of human life. However, in the face of human participation, whether pedestrians or users are encountered, the service robot needs to not only complete the route selection in the moving process, but also be aware of human causes, and fully consider the planning constraint brought by some human participation.
The current research usually uses a cost function method, a social force model method or a learning-based method to achieve the purpose of considering the constraint of the personnel. The method is based on learning or semantics and the like, uses features to capture important attributes of human behaviors, and estimates feature weights through machine learning. Once a function associated with a person is defined, the characteristics of the conventional cost function method are not changed, and therefore, further improvement in simulating dynamic uncertainty and integrating person movement is required.
Although the human-computer coexistence scene has complexity, the most basic behavior of the robot is required to be trend and avoidance, and how to reasonably describe the personnel requirements directly influences the intelligent level of the service robot. In addition, although the global and local combined path planning can satisfy the mobility, how to better configure the path planning greatly influences the path generation quality, for example, a more active planning algorithm can obviously reduce the local obstacle avoidance burden, and the mobile robot has better motion real-time performance and is convenient for motion control.
Aiming at the problem of failure instability under the influence of participation of personnel in the existing service robot path planning technology, a more targeted solution is not provided at present.
Disclosure of Invention
The invention designs a service robot navigation planning method for distinguishing personnel influence, which aims to solve the problems in the prior art.
The technical scheme of the invention is as follows:
a service robot navigation planning method for distinguishing personnel influence comprises the following steps:
first part, cost modeling
And 2, quantifying abstract description of influence of pedestrians and users. Modeling to obtain personnel impact cost C1. (quantitative analysis)
And 3, automatically and synchronously positioning and establishing a map by the service robot to obtain a static environment map. Output environment map cost C2Representing the service robot passage cost.
Step 4, obtaining a total cost map CGeneral assembly＝C1+C2. And superposing the personnel influence cost and the static environment map cost.
Second part, sensing handover
And 5, detecting the participation of the personnel by the sensor. The dynamic personnel are detected from the environment, and then the user and the irrelevant pedestrian are further distinguished.
And 6, determining and switching the target position. And (5) determining the target position and whether to switch the thread according to whether the user is found in the step 5.
Third, navigation planning
And 7, starting the global path planner, optimizing a heuristic function and obtaining a global path to the target point.
And 8, starting a local planner, optimizing a track prediction scoring mechanism, and selecting the optimal planning speed to avoid obstacles for advancing.
The invention has the beneficial effects that:
(1) based on the fact that the service robot works in a person participation scene frequently, characteristic classification and further quantification are carried out on the additional influence of the person, and the additional influence is from abstract to concrete;
(2) the influence model is brought into an environment map cost model, and is closely linked with a combined navigation planning process, and the consideration factors are comprehensive;
(3) and on the perception level, whether the personnel participate is checked, users and pedestrians are distinguished, the navigation target position is adjusted, and avoidance or approach is performed according to the approachability.
(4) The invention solves the problem of failure and instability under the influence of participation of personnel in the existing service robot path planning technology, thereby greatly improving the user experience and the execution capacity of the service robot.
Drawings
Fig. 1 is a flowchart of a service robot navigation planning method under the influence of human participation according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a user and pedestrian quantitative modeling method according to an embodiment of the invention. The system comprises a pedestrian and user interaction model, a two-dimensional non-elevation Gaussian function modeling schematic diagram and a two-dimensional;
FIG. 3 is a schematic diagram of a combined path plan according to an embodiment of the invention;
fig. 4 is a schematic diagram illustrating the effect of human participation in influencing the navigation effect of the service robot in a specific scenario according to an embodiment of the present invention. Wherein, (a) and (b) are respectively pedestrian and user.
Detailed Description
In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention.
The invention provides a service robot navigation planning method for distinguishing personnel influence. FIG. 1 is a flow chart of a technical method according to an embodiment of the invention. As shown in fig. 1, the method includes:
the first part, cost modeling.
And 2, quantifying abstract description of influence of pedestrians and users. Respectively using two-dimensional non-elevation gaussian functions to carry out geometric modeling to obtain personnel influence cost C1. (quantitative analysis)
Further, the human influence quantification strategy in step 2 is detailed as follows:
(1) the approach and avoidance characteristics are visually represented as approaching and departing of the mobile robot in distance, namely distance (or space) language is used as a quantification standard.
(2) The pedestrian is strange to the service robot, and intrusion into the private space of the strange object is not considered reasonable. Therefore, a two-dimensional non-elevation Gaussian function is used for distance and geometric modeling, the influence of the pedestrian is converted into distance corresponding values of various directions and various levels through the size of a Gaussian curve value and cutoff processing, the curve value is distributed in a [0,1] interval, and the higher the value is, the closer the value is. The user is a service object of the service robot and has intimacy, so that the reverse two-dimensional non-elevation Gaussian function geometric modeling is established (namely the whole is negative), the curve values are distributed in the range of [ -1,0], and the lower the value is, the closer the value is.
(3) The geometric modeling and the two-dimensional non-elevation Gaussian function are designed as follows:
the formula (1) is a Gaussian function formula, f (x, y) is a quantized cost value, (x0, y0) is a position of a person, and the amplitude A is 1 or-1 according to the role of the person; as the sight range of the personnel and the principle of driving towards the right are more sensitive to the approaching towards the front than the rear and the approaching towards the right than the left, the Gaussian function takes different variance values sigma + x, sigma-x, sigma + y and sigma-y in different directions according to the orientation of the personnel as the front.
And step 3, obtaining a static environment map. The service robot senses a service place and static obstacles by utilizing a sensor to perform an SLAM process and outputs an environment map cost C2. The map is represented in the form of a grid, the grid occupying the stateThe value represents the service robot passing cost.
Step 4, obtaining a total cost map CGeneral assembly＝C1+C2. And superposing the personnel influence cost and the static environment map cost.
Further, the total cost map obtaining strategy described in step 4 is detailed as follows:
(1) in the SLAM (Simultaneous Localization and mapping) process of the service robot, Hokuyo laser radar and Rao-Blackwellized particle filter algorithm are utilized, odometry information provided by the robot body is added, and the obtained environment map cost C is output2. The environment map is represented in the form of a grid occupying a binary map. An occupancy value of 1 indicates that the grid has an obstacle present, and an occupancy value of 0 indicates that free passage is possible.
(2) Model C of the cost of the personnel impact1Static environment map cost C2The two parts are superposed. As shown in formula (2), due to C1The part with the original occupancy value of 1 is not likely to have any more people present, and the grid cell values representing the movement cost values are finally distributed to-1, 1]An interval. In this way, the established personnel influence cost model is taken into consideration to restrict the behavior of the service robot so as to make the service robot have tendency and evasion.
CGeneral assembly＝C1+C2(2)
And a second part, sensing handover.
And 5, detecting the participation of the personnel by the sensor. The dynamic personnel are detected from the environment, and then the user and the irrelevant pedestrian are further distinguished. The method comprises the following steps:
the method comprises the steps of clustering and segmenting the point cloud of the human legs, taking laser scanning as input, detecting data clusters by using a random forest classifier, taking the data clusters as human leg candidates, and detecting dynamic personnel from the environment. Further, the depth camera adopts a template matching method to identify the outline shape of the dynamic personnel, and compares the outline shape with a user database to distinguish the user from irrelevant pedestrians.
And 6, determining and switching the target position. And 5, temporarily searching the target position to find the user in the advancing process when the user is not detected. And when the sensor detects the user, the thread is switched to take the user as a target position.
And a third part, navigation planning.
And 7, starting the global path planner to obtain a global path to the target point. After obtaining the input of the target location, at the total cost map CGeneral assemblyThe improved global planning algorithm is used, the heuristic function part is added with personnel cost, and path searching is carried out based on the cost of the grid map.
Further, the global planning strategy in step 7 is detailed as follows:
(1) the map server receives cost map information; once the input of the target position is obtained, observing by using a sensor, and obtaining the current position and the current pose orientation of the service robot based on Monte Carlo autonomous positioning; the global planner updates the position of the user relative to the robot according to the planning frequency, and the personnel influence cost is also updated at the stage to register the user information.
(2) Global Path planner Start at Total cost map CGeneral assemblyAnd performing path search. With the eight-connected grid as a search strategy, the traditional a-x algorithm is optimized and improved here, and the heuristic function h part of the algorithm is initially the distance estimation from the current position of the service robot to the target position. On this basis, h is also modified to be the sum of the distance cost and the personnel impact cost, and the dynamic cost is also weighted to speed up the search process. And finally obtaining the global route. As shown in formula (3).
f＝g+h+δh1(3)
Wherein f is a movement cost function of the service robot from the initial position to the target position, g is an actual consumption cost from the initial point to the current point, h is a distance estimation from the current position to the target position, and h is a distance estimation1And d is a weighting coefficient, and a passing value is taken.
And 8, starting the local planner to avoid obstacles. And (3) enabling a local algorithm to follow the global path, optimizing a track prediction scoring mechanism, observing in real time by using a sensor to avoid local obstacles, selecting an optimal planning speed and sending the optimal planning speed to an execution mechanism.
Further, the local planning strategy in step 8 is detailed as follows:
(1) issuing a global route as a reference path to a local planner; meanwhile, the sensor performs real-time observation, scans barriers and personnel detection information.
(2) Sampling the group (linear velocity v, angular velocity w) in a speed space formed by the maximum and minimum speed limit, the motor torque limit and the parking distance limit; forward simulating a trajectory according to the velocity set and the planned time interval Δ T; scoring the trajectory, scoring terms: the size of a target orientation angle, the size of a speed, the distance between the target orientation angle and a nearest obstacle, and the size of the shortest distance between the target orientation angle, the speed and the nearest obstacle optimized by the method are added; selecting the planning speed (v) corresponding to the simulation track with the highest score0，w0)。
(3) Finally, directly mixing (v)0，w0) And sending the speed to a mobile chassis, and performing speed smooth output by an executing mechanism.
Claims (1)
1. A service robot navigation planning method for distinguishing personnel influence is characterized by comprising the following steps:
first part, cost modeling
Step 1, personnel influence characteristic classification: according to different requirements of the service robot, people are divided into two types: pedestrians and users, namely the users are served objects; the most basic behavior characteristics of the service robot are required to be expressed as approach and avoidance;
step 2, quantifying abstract description of influence of pedestrians and users; respectively using two-dimensional non-elevation gaussian functions to carry out geometric modeling to obtain personnel influence cost C1；
(1) The behavior approaching and avoiding characteristics are visually expressed as approaching and departing of the service robot in distance, namely distance language is used as a quantification standard;
(2) pedestrians are strange to the service robot, and intrusion into the private space of the strange object is considered unreasonable; therefore, a two-dimensional non-elevation Gaussian function is used for distance and geometric modeling, the influence of the pedestrian is converted into distance corresponding values of various levels in various directions through the size of a Gaussian curve value and truncation processing, the curve value is distributed in a [0,1] interval, and the higher the curve value is, the closer the curve value is; the user is a service object of the service robot and has intimacy, so that a reverse two-dimensional non-elevation Gaussian function is established for geometric modeling, namely the whole is negative, the curve value is distributed in the range of [ -1,0], and the lower the value is, the closer the value is;
(2) the geometric modeling and two-dimensional non-elevation gaussian functions are:
in the formula (1), f (x, y) is a quantization cost value, (x)0,y0) The amplitude A is 1 or-1 according to the role of the person for the position of the person; as the visual field range of personnel and the principle of driving towards the right are more sensitive to the front and the back of the approaching vehicle and more sensitive to the right and the left of the approaching vehicle, the Gaussian function takes different variance values sigma + x, sigma-x, sigma + y and sigma-y in different directions according to the direction of the personnel as the front;
step 3, obtaining a static environment map, sensing a service place and a static obstacle by the service robot in a SLAM (Simultaneous localization and mapping) process by using a sensor, and outputting a static environment map cost C2(ii) a The map is represented in a grid form, and the grid occupation state value represents the passing cost of the service robot;
step 4, obtaining a total cost map CGeneral assembly＝C1+C2(ii) a Superposing two parts of personnel influence cost and static environment map cost;
(1) in the SLAM process of the service robot, the environment map cost C is obtained by utilizing the Hokuyo laser radar and the Rao-Blackwellized particle filter algorithm and additionally adding the odometer information provided by the service robot body and outputting2(ii) a The environment map is represented in a grid form of an occupancy binary map, the occupancy value of 1 represents that the grid has obstacles, and the occupancy value of 0 represents that the environment map can freely pass through;
(2) influencing personnel by the cost C1Static environment map cost C2Two parts are superimposed, as in formula (2), with cost C due to personnel impact1The part with the original occupation value of 1 can not be used any moreIf a person is present, the grid cell values representing the movement cost values are finally distributed to [ -1,1]An interval; therefore, the cost of the influence of the established personnel is taken into consideration to restrict the behavior of the service robot so as to make the service robot have tendency and evasion;
Cgeneral assembly＝C1+C2(2)
Second part, sensing handover
And 5, detecting personnel participation by a sensor: detect dynamic personnel from the environment earlier, further distinguish user and irrelevant pedestrian again, it is specific: performing point cloud cluster segmentation on human legs, taking laser scanning as input, detecting data clusters by using a random forest classifier, taking the data clusters as human leg candidates, and detecting dynamic personnel from the environment; further, the depth camera adopts a template matching method to identify the outline shape of the dynamic personnel, compares the outline shape with a user database and distinguishes users and pedestrians;
step 6, determining and switching the target position: step 5, when the user is not detected, temporarily searching the target position to search the user in the process of advancing; when the sensor detects a user, the thread is switched, and the user is taken as a target position;
third, navigation planning
Step 7, starting the global path planner to obtain a global path to a target point; after obtaining the input of the target location, at the total cost map CGeneral assemblyAn improved global planning algorithm is used, the heuristic function is added with personnel cost, and path searching is carried out based on the cost of the grid map;
the map server receives cost map information; once the input of the target position is obtained, observing by using a sensor, and obtaining the current position and the current pose orientation of the service robot based on Monte Carlo autonomous positioning; the global path planner updates the position of the user relative to the service robot according to the planning frequency, the personnel influence cost is also updated at the stage, and the user information is registered;
(1) global Path planner Start at Total cost map CGeneral assemblyPerforming path search; the eight-connected grid is used as a search strategy, the traditional A-x algorithm is optimized and improved at the place, and the heuristic function of the traditional A-x algorithmThe number h part is initially the distance estimation from the current position of the service robot to the target position; on the basis, h is modified into the sum of the distance cost and the personnel influence cost, and dynamic cost is weighted to accelerate the search process, so that a global route is finally obtained, as shown in formula (3):
f＝g+h+δh1(3)
wherein f is a movement cost function of the service robot from the initial position to the target position, g is an actual consumption cost from the initial point to the current point, h is a distance estimation from the current position to the target position, and h is a distance estimation1Taking a pass-by value as the personnel influence cost of the grid and delta as a weighting coefficient;
step 8, starting a local planner and advancing in an obstacle avoidance manner; enabling a local algorithm to follow the global path, optimizing a track prediction scoring mechanism, utilizing a sensor to observe in real time to avoid local obstacles, and selecting an optimal planning speed to send to an execution mechanism;
(1) the global route issues a reference path to a local planner; meanwhile, the sensor performs real-time observation to scan obstacles and personnel detection information;
(2) sampling the group (linear velocity v, angular velocity w) in a speed space formed by the maximum and minimum speed limit, the motor torque limit and the parking distance limit; forward simulating the trajectory according to the set of (linear velocity v, angular velocity w) and the planned time interval Δ T; scoring the trajectory, scoring terms: the target orientation angle, the speed, the distance to the nearest obstacle and the shortest distance to the global path; selecting the planning speed (v) corresponding to the simulation track with the highest score0，w0)；
(3) Finally, directly mixing (v)0，w0) And sending the speed to a mobile chassis, and performing speed smooth output by an executing mechanism.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911211423.7A CN110906946A (en) | 2019-12-02 | 2019-12-02 | Service robot navigation planning method for distinguishing personnel influence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911211423.7A CN110906946A (en) | 2019-12-02 | 2019-12-02 | Service robot navigation planning method for distinguishing personnel influence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110906946A true CN110906946A (en) | 2020-03-24 |
Family
ID=69821446
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911211423.7A Pending CN110906946A (en) | 2019-12-02 | 2019-12-02 | Service robot navigation planning method for distinguishing personnel influence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110906946A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112556686A (en) * | 2020-12-08 | 2021-03-26 | 中国人民解放军61618部队 | Shortest time path planning method capable of predicting dynamic space-time environment |
| WO2021248857A1 (en) * | 2020-06-08 | 2021-12-16 | 特斯联科技集团有限公司 | Obstacle attribute discrimination method and system, and intelligent robot |
| CN115200588A (en) * | 2022-09-14 | 2022-10-18 | 煤炭科学研究总院有限公司 | SLAM autonomous navigation method and device for mobile robot |
| CN117470253A (en) * | 2023-12-28 | 2024-01-30 | 中国人民解放军国防科技大学 | Tensor field-based robot path planning method, device, equipment and medium |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107272680A (en) * | 2017-06-16 | 2017-10-20 | 华南理工大学 | A kind of automatic follower method of robot based on ROS robot operating systems |
-
2019
- 2019-12-02 CN CN201911211423.7A patent/CN110906946A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107272680A (en) * | 2017-06-16 | 2017-10-20 | 华南理工大学 | A kind of automatic follower method of robot based on ROS robot operating systems |
Non-Patent Citations (2)
| Title |
|---|
| 陈伟华: "社会环境的轮式移动机器人定位导航方法研究", 《中国博士学位论文全文数据库信息科技辑》 * |
| 陈赢峰: "大规模复杂场景下室内服务机器人导航的研究", 《中国博士学位论文全文数据库信息科技辑》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021248857A1 (en) * | 2020-06-08 | 2021-12-16 | 特斯联科技集团有限公司 | Obstacle attribute discrimination method and system, and intelligent robot |
| CN112556686A (en) * | 2020-12-08 | 2021-03-26 | 中国人民解放军61618部队 | Shortest time path planning method capable of predicting dynamic space-time environment |
| CN112556686B (en) * | 2020-12-08 | 2023-04-11 | 中国人民解放军61618部队 | Shortest time path planning method capable of predicting dynamic space-time environment |
| CN115200588A (en) * | 2022-09-14 | 2022-10-18 | 煤炭科学研究总院有限公司 | SLAM autonomous navigation method and device for mobile robot |
| CN117470253A (en) * | 2023-12-28 | 2024-01-30 | 中国人民解放军国防科技大学 | Tensor field-based robot path planning method, device, equipment and medium |
| CN117470253B (en) * | 2023-12-28 | 2024-03-22 | 中国人民解放军国防科技大学 | Tensor field-based robot path planning method, device, equipment and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110906946A (en) | Service robot navigation planning method for distinguishing personnel influence | |
| US10696300B2 (en) | Vehicle tracking | |
| Yao et al. | On-road vehicle trajectory collection and scene-based lane change analysis: Part II | |
| Laugier et al. | Probabilistic analysis of dynamic scenes and collision risks assessment to improve driving safety | |
| CN114384920A (en) | Dynamic obstacle avoidance method based on real-time construction of local grid map | |
| Wojke et al. | Moving vehicle detection and tracking in unstructured environments | |
| Habibi et al. | Context-aware pedestrian motion prediction in urban intersections | |
| Ortega et al. | Overtaking maneuver scenario building for autonomous vehicles with PreScan software | |
| Gwak et al. | A review of intelligent self-driving vehicle software research | |
| CN116611603B (en) | Vehicle path scheduling method, device, computer and storage medium | |
| Chavez-Garcia | Multiple sensor fusion for detection, classification and tracking of moving objects in driving environments | |
| JP2024014875A (en) | Object localization for autonomous driving by visual tracking and image reprojection | |
| CN116573017A (en) | Urban rail train running clearance foreign matter sensing method, system, device and medium | |
| US20230311932A1 (en) | Merging object and background radar data for autonomous driving simulations | |
| Qing et al. | A novel particle filter implementation for a multiple-vehicle detection and tracking system using tail light segmentation | |
| Chen et al. | Design and Implementation of AMR Robot Based on RGBD, VSLAM and SLAM | |
| CN113741550B (en) | Mobile robot following method and system | |
| Chen et al. | Costmap generation based on dynamic obstacle detection and velocity obstacle estimation for autonomous mobile robot | |
| WO2023192397A1 (en) | Capturing and simulating radar data for autonomous driving systems | |
| CN113298044B (en) | Obstacle detection method, system, device and storage medium based on positioning compensation | |
| Sun et al. | Detection and state estimation of moving objects on a moving base for indoor navigation | |
| Han et al. | Novel cartographer using an oak-d smart camera for indoor robots location and navigation | |
| Lee et al. | Visually-extended vector polar histogram applied to robot route navigation | |
| Sanberg et al. | From stixels to asteroids: Towards a collision warning system using stereo vision | |
| Zhang | An improved DBSCAN Algorithm for hazard recognition of obstacles in unmanned scenes |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200324 |
|
| RJ01 | Rejection of invention patent application after publication |