CN116225004A - Obstacle avoidance method for six-wheel independent driving independent steering robot - Google Patents
Obstacle avoidance method for six-wheel independent driving independent steering robot Download PDFInfo
- Publication number
- CN116225004A CN116225004A CN202310103879.1A CN202310103879A CN116225004A CN 116225004 A CN116225004 A CN 116225004A CN 202310103879 A CN202310103879 A CN 202310103879A CN 116225004 A CN116225004 A CN 116225004A
- Authority
- CN
- China
- Prior art keywords
- obstacle
- robot
- obstacle avoidance
- track
- distance
- 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
- 238000000034 method Methods 0.000 title claims abstract description 67
- 230000008569 process Effects 0.000 claims abstract description 33
- 230000003068 static effect Effects 0.000 claims abstract description 10
- 238000009499 grossing Methods 0.000 claims abstract description 4
- 230000033001 locomotion Effects 0.000 claims description 61
- 230000004888 barrier function Effects 0.000 claims description 16
- 230000001133 acceleration Effects 0.000 claims description 7
- 238000005457 optimization Methods 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000013459 approach Methods 0.000 claims description 5
- 239000011159 matrix material Substances 0.000 claims description 5
- 230000002829 reductive effect Effects 0.000 claims description 5
- 230000009471 action Effects 0.000 claims description 4
- 230000000452 restraining effect Effects 0.000 claims description 4
- 101100039010 Caenorhabditis elegans dis-3 gene Proteins 0.000 claims description 3
- 101100030351 Schizosaccharomyces pombe (strain 972 / ATCC 24843) dis2 gene Proteins 0.000 claims description 3
- 238000013461 design Methods 0.000 claims description 3
- 238000006073 displacement reaction Methods 0.000 claims description 3
- 238000011160 research Methods 0.000 description 12
- 239000004973 liquid crystal related substance Substances 0.000 description 6
- 238000005070 sampling Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000036961 partial effect Effects 0.000 description 2
- 239000004566 building material Substances 0.000 description 1
- 230000009194 climbing Effects 0.000 description 1
- 238000011217 control strategy Methods 0.000 description 1
- 230000007123 defense Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 230000000670 limiting effect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
- G05D1/0219—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory ensuring the processing of the whole working surface
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Aviation & Aerospace Engineering (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
The obstacle avoidance method of the six-wheel independent driving independent steering robot comprises the steps of solving an obstacle avoidance path and an obstacle avoidance strategy based on fuzzy decision; after the sensor detects the obstacle, the obstacle information is sent to the obstacle avoidance controller and the obstacle avoidance planner, then the optimal path point is solved through the constraint of the actual environment, the optimal path is obtained through track smoothing and polynomial fitting, the optimal path is output after being scattered into coordinate points, the obstacle avoidance strategy is selected after the coordinate points are output, the obstacle avoidance is carried out according to the strategy, and the whole obstacle avoidance process is finished. The invention solves the obstacle avoidance problem of six-wheel independent driving independent steering robots under the condition of encountering static obstacles and dynamic obstacles in the track tracking process, and verifies the robot aiming at actual conditions and physical models.
Description
Technical Field
The invention belongs to the technical field of motion control of wheeled mobile robots, and particularly relates to an obstacle avoidance method and an obstacle avoidance strategy of a wheeled robot.
Background
With the development of intelligent control technology, mobile robots are increasingly used in industrial production, military national defense, teaching and scientific research and other fields. The mobile robot is mainly divided into a crawler type mobile robot, a leg-foot type mobile robot, a wheel type mobile robot and the like, and compared with the crawler type mobile robot and the leg-foot type mobile robot, the wheel type mobile robot is flexible in movement, simple in structure, low in control difficulty and higher in bearing capacity, so that the movement control research of the wheel type mobile robot is one of important directions of future intelligent robot research.
Path planning and trajectory tracking of robots are important issues in research for mobile robot motion control. However, in the planning and tracking process, the occurrence of static barriers or dynamic barriers is a very easy occurrence condition in actual control, and the autonomous obstacle avoidance capability is an important index of the intelligent degree of the mobile robot, and is also an important guarantee for the mobile robot to stably, safely and efficiently complete tasks in a complex environment. Therefore, research on intelligent obstacle avoidance of robots is also always one of the hot problems of research.
The six-wheel independent driving independent steering robot is a new wheel type robot model, and compared with the traditional wheel type mobile robot, the six-wheel independent driving independent steering robot has certain progress in the aspects of single body loading capacity, climbing capacity, driving capacity and the like. Therefore, the six-wheel independent driving independent steering robot has stronger adaptability to various terrains, and has important significance for transporting goods, rescue materials, building materials and the like in special environments. Besides, the six-wheel independent driving independent steering robot is driven by a motor, so that noise is greatly reduced, the number of driving wheels can be adjusted according to working conditions, and the cruising ability is improved. Because of the great enhancement of the load capacity, mechanical arms and the like can be added on the platform to process tasks under different conditions. Moreover, the system has high flexibility and high mobility, can adapt to various limit working conditions, and has high applicability in rescue and relief work, geological investigation and other special fields.
Fuzzy decision refers to mathematical theory and method of making decisions in a fuzzy environment. Strictly speaking, the real decisions are mostly fuzzy decisions. The study of fuzzy decision starts later, but the area involved is very broad and has not been clearly defined. Common fuzzy decision methods include fuzzy sequencing, fuzzy optimizing, fuzzy countermeasures and the like. Fuzzy optimization mainly refers to finding the optimal scheme after a scheme set and various objective functions and limiting conditions are given. If the objective function or constraint is fuzzy, then the optimization is referred to as fuzzy optimization. One approach to objective function blurring is to find conditional extremum by analysis and operation of the fuzzy number with the fuzzy number as the objective function value. Fuzzification of constraints is the definition of constraints as a fuzzy set. Such popularization in linear programming results in research of fuzzy linear programming, and as a result, the application range of the common linear programming is wider, and the method can be more flexibly adapted to various different conditions.
The obstacle avoidance strategy mainly aims at avoiding in real time according to obstacle information in the environment after the robot detects the obstacle in the environment through the sensor and generates an obstacle avoidance path. The movement of the obstacle in the actual environment is usually complex, so that adjustment is required at every moment in the future after the obstacle avoidance path is generated. The obstacle avoidance strategy is very important research content in the intelligent obstacle avoidance process of the robot, and the selection of the proper strategy directly influences the obstacle avoidance effect of the robot.
Current obstacle avoidance studies are mainly directed to situations where obstacles appear in the direction of travel of the robot, and no other situations are analyzed. Therefore, the method analyzes the situation that the obstacle appears in each direction in the movement process of the robot, and has great significance in the aspect of intelligent robot obstacle avoidance research in the aspects of generating a proper obstacle avoidance path and an effective obstacle avoidance strategy.
Disclosure of Invention
Aiming at the problems of the existing obstacle avoidance technology of the six-wheel independent driving independent steering robot in the track tracking process, the invention provides an obstacle avoidance method of the six-wheel independent driving independent steering robot, so as to solve the obstacle avoidance path and the obstacle avoidance strategy according to different situations of obstacles, and solve the obstacle avoidance problem when different dynamic/static obstacles appear in the movement process.
The invention relates to an obstacle avoidance method for a six-wheel independent driving independent steering robot, which adopts the following scheme:
solving an obstacle avoidance path and an obstacle avoidance strategy based on fuzzy decision; after the sensor detects the obstacle, the obstacle information is sent to the obstacle avoidance controller and the obstacle avoidance planner, then the optimal path point is solved through the constraint of the actual environment (distance constraint, control quantity constraint and the like), the optimal path is obtained through track smoothing and polynomial fitting, the optimal path is scattered into a coordinate point and then is output, the obstacle avoidance strategy is selected after the coordinate point is output, the obstacle avoidance is carried out according to the strategy, and the whole obstacle avoidance process is finished.
The process for solving the obstacle avoidance path is as follows:
in the process of planning an obstacle avoidance path, performing expansion processing according to the size of the robot and the size of the obstacle, and expanding the obstacle according to the obstacle, the necessary safety margin of the robot and the information of the obstacle, wherein the expansion range of the robot is added into the part (in the process of planning the obstacle avoidance path) because of the safety margin; taking the situation that the obstacle is large and the robot possibly passes through the middle of the obstacle into consideration, dividing the obstacle, and setting an obstacle dividing point at the outline of the obstacle to perform dividing;
in the solving process of the obstacle avoidance path, designing an obstacle avoidance penalty function and a cost function of a design deviation reference track; the obstacle avoidance penalty function is used for solving and restraining the obstacle avoidance function, the size of the function is adjusted through the distance deviation between the robot mass center coordinates and the obstacle dividing points, and the function value is larger as the distance is closer; taking a cost function of the robot deviated from the reference track as a constraint condition for solving;
solving the obstacle avoidance path directly obtains the coordinates of the optimal path point, so that definition is that:
ξ＝[x a y a ] T ，
wherein x is a y a Is a coordinate point of the obstacle avoidance path.
In order to improve the accuracy of obstacle avoidance, a distance penalty function D is defined obs Adding soft constraint into obstacle avoidance function
K is a very large positive integer, which is aimed at the situation that a dynamic obstacle suddenly approaches the robot in the moving process of the robot, when the distance between the robot and the obstacle is relatively short, the soft constraint item is relatively large, and the weight influence is large; when the distance between the robot and the obstacle is far, the soft constraint item is smaller, and the weight influence is small;
according to the punishment items and constraints, designing a barrier avoidance planner based on a mathematical optimization method, wherein the barrier avoidance planner is designed as follows:
wherein: p is a weight matrix;
the term is the control quantity of the MPC, and the action in the obstacle avoidance planner is to restrict the obstacle avoidance path, so that the obstacle avoidance path obtained by the planner is ensured to conform to the restriction of the robot kinematic model; the result xi solved by the obstacle avoidance planner is the obstacle avoidance path of the robot; p (P) obs To avoid barrier penalty function D obs Distance penalty function. In the constraint condition, a->
And the MPC control quantity is ensured to be in accordance with a motion model of the robot, and gamma is a cost function of the deviation reference track.
And processing the obstacle avoidance path solved by the obstacle avoidance planner through curve fitting, wherein the curve fitting generates the smoothed track points into track points which conform to the robot constraint (such as continuous robot position requirement, yaw angle requirement is a first-order continuous curve, acceleration constraint requirement is a second-order continuous curve and the like).
The obstacle avoidance penalty function P obs The following are provided:
wherein p is obs Is the barrier avoidance penalty function weight coefficient, (x) i ，y i ) Is the coordinates of the dividing points of the barrier, and E is an extremely small positive number, so that the situation that 0 occurs in the denominator is avoided;
the cost function of the deviation reference track is as follows:
wherein x is min ，x max ，y min ，y max Is the global map coordinate range, x r ，y r Is the reference track coordinate, x a ，y a Is a coordinate point of the obstacle avoidance path.
The distance penalty function is:
wherein x is obs ，y obs For the heart coordinate of the obstacle, x rob ，y rob Is the robot mass heart coordinate d obs Weighting coefficients are distance penalty functions.
And the curve fitting is carried out by selecting n times of polynomial fitting as a fitting curve and carrying out fitting based on the least square method principle, wherein the fitting function is as follows:
X＝a 1 x n +a 2 x n-1 +a 3 x n-2 +…+a n x+a n+1 ，
Y＝b 1 y n +b 2 y n-1 +b 3 y n-2 +…+b n y+b n+1 ，
wherein a= [ a ] 1 ，a 2 …a n+1 ]，b＝[b 1 ，b 2 …b n+1 ]Polynomial X and Y coefficients, respectively; n=4 is chosen, i.e. the order of the fit is 4 times, and X and Y are the coordinates of the curve after the fit.
The obstacle avoidance strategy based on fuzzy decision is as follows:
the robot obstacle avoidance strategy is selected by predicting an obstacle track and a robot track, predicting whether collision occurs or not, and then performing strategy selection; when the static obstacle is regarded as the condition that the motion speed and the acceleration are 0 and the safety distance threshold value is exceeded, the obstacle avoidance track is planned, and track tracking is performed after the obstacle avoidance track is planned;
For a dynamic obstacle, if the moving track of the obstacle is a straight line, directly judging whether collision occurs after a certain time or not according to the calculation result of a threshold function, and then solving an optimal path by utilizing constraint; if the moving track of the obstacle is a curve, predicting the track after a certain time according to the differential function of the displacement of the obstacle to the time, judging the movement condition of the robot under the time according to the movement time when the robot intersects with the track of the robot, judging whether collision occurs or not by utilizing the centroid distance of the robot and the obstacle, and then solving the optimal path.
An asymmetric model in a fuzzy decision is adopted in the selection of the robot obstacle avoidance decision, the obstacle avoidance condition of the robot is taken as a constraint condition, and the selection of the robot obstacle avoidance strategy is taken as an objective function; in an asymmetric model, taking acceptance constraints as preconditions, the position of both the target and the constraints is not symmetric; giving an objective function f (X) on a domain X and a constraint condition fuzzy set D on X, wherein an optimal solution m for maximizing f under the constraint D is a fuzzy subset on X, the fuzzy subset has a membership function, when the set on the right end of an equation is an empty set, mum (X) is equal to 0, the given domain X is a set of all conditions, the current condition encountered by the robot is the constraint D, and the objective function is an obstacle avoidance strategy of the robot;
The robot solves the objective function by taking the environment information and the obstacle information when encountering the obstacle as constraint conditions, and selects a proper obstacle avoidance strategy.
The dynamic barrier is specifically divided into the following three cases:
(1) In the first case, the movement direction of the obstacle is the same as the movement direction of the robot track;
at the moment, the planning of the obstacle avoidance track is to enable the robot to select an optimal track from the left side or the right side of the obstacle to overrun, and return to the original track for tracking after overrun; after the obstacle avoidance track is re-planned, outputting new track points to a track tracking controller, and calculating the control quantity of the robot according to an MPC algorithm to track; when the position of the robot mass center and the position of the obstacle mass center are positioned on the same horizontal line, adding speed feedback proportional control into a speed item to enable the robot to perform obvious overtaking operation and return to an original track as soon as possible, and adding differential control to prevent overshoot;
wherein v is mpc The speed, k of the robot control quantity calculated by the MPC track tracking controller p1 And k d1 Is a speed feedback proportional control and differential control coefficient, e 11 Is the deviation value t of the obstacle avoidance track of the robot from the original track k Is the current moment of the robot movement, t d1max Is the robot moves to d 1max Position moment, t d1min Is d 1min The time of the position is shown;
(2) In the second case, the movement direction of the obstacle intersects with the movement direction of the robot, and the movement speed of the obstacle is higher;
the situation is that whether the robot tracking track and the obstacle moving track collide at the intersection point or not is judged, and the judgment basis is related to the moving speed of the robot and the obstacle; if the moving speed of the obstacle is relatively high, the robot is properly decelerated before reaching the collision point, then the robot overruns from the rear of the obstacle after the obstacle passes over the intersection point, and then returns to the original track for tracking; the distance between the robot and the obstacle is small and then large, the speed needs to be reduced when the distance is small, the speed feedback proportional control is added on the basis of the MPC control quantity, the MPC control is restored when the distance is large, and meanwhile the differential control is added to prevent overshoot when the original track is tracked;
wherein v is mpc The speed, k of the robot control quantity calculated by MPC p2 And k d2 Is a speed feedback proportional control and differential control coefficient, e 21 Is the deviation of the obstacle avoidance track of the robot from the original track, e 2 Is the distance d between the position of the robot mass center and the position of the obstacle mass center dis2 Is the absolute distance between the mass center of the robot and the mass center of the obstacle in the x direction, d 2min And d 2max The distance threshold value of the mass center of the robot and the obstacle is respectively;
(3) In the third case, the movement direction of the obstacle intersects with the movement direction of the robot, and the movement speed of the obstacle is slower;
the robot is accelerated properly before reaching the collision point, and overruns from the front of the obstacle when the obstacle does not cross the intersection point, and then returns to the original track for tracking; the distance between the robot and the obstacle is small and then large, the speed needs to be increased when the distance is small, the speed feedback proportional control is added on the basis of the MPC control quantity, the obstacle avoidance is accelerated, the MPC control is restored when the distance is large, and meanwhile the differential control is added to prevent overshoot when the original track is tracked;
wherein v is mpc The speed, k of the robot control quantity calculated by MPC p3 And k d3 Is a speed feedback proportional control and differential control coefficient, e 31 Is the deviation of the obstacle avoidance track of the robot from the original track, e 3 Is the distance d between the position of the robot mass center and the position of the obstacle mass center dis3 Is the absolute distance between the mass center of the robot and the mass center of the obstacle in the x direction, d 3min And d 3max The distance threshold value of the robot mass center and the obstacle is respectively.
The obstacle avoidance method and the obstacle avoidance strategy provided by the invention solve the obstacle avoidance problem of six-wheel independent driving independent steering robots under the condition of encountering static obstacles and dynamic obstacles in the track tracking process. And aiming at the actual situation and the physical model, the invention content provided by the invention is verified on the robot. Besides, the invention expands various different situations and new robot models of the robot in the obstacle avoidance process, and provides a certain reference for other obstacle avoidance methods.
Compared with the prior art, the invention has the following characteristics:
1. the invention provides obstacle avoidance conditions under various obstacle conditions, and provides a corresponding obstacle avoidance method and an obstacle avoidance strategy. After the obstacle avoidance path is planned by combining the movement condition of the obstacle, the obstacle avoidance is performed through the adjustment of the speed of the robot and the selection comprehensive control of a proper obstacle avoidance strategy, so that the obstacle avoidance method meets various complex situations in the actual situation, and the provided scheme can also solve various obstacle avoidance problems of different types.
2. The invention uses six-wheel independent driving independent steering model. Six-wheel independent driving independent steering robots have less research content in the research of the wheel type mobile robots at the present stage due to the complicated control characteristics, larger control quantity calculation compared with two-wheel and four-wheel robots and the like. When the obstacle avoidance path is planned, the six-wheeled robot has larger volume, and the steering condition and the path condition of the robot are required to be simultaneously considered for comprehensive control. The motion state of the robot is judged through the global positioning system and each wheel axle encoder of the robot, and the accuracy of the motion of the robot is ensured by utilizing the single wheel speed and steering fine control.
3. The obstacle avoidance strategy of the robot is selected by adopting a fuzzy decision method, various obstacle avoidance problems encountered by the robot are taken as constraints, the obstacle avoidance strategy of the robot is taken as an objective function, the problem of obstacle avoidance under the problem of different types of obstacles is solved by utilizing an asymmetric model in fuzzy optimization, the operation difficulty is reduced, and the obstacle avoidance reliability is ensured.
Drawings
FIG. 1 is a schematic diagram of an obstacle avoidance process according to the present invention.
Fig. 2 is a diagram of obstacle inflation segmentation.
Fig. 3 is a schematic representation of partial segment trajectory fitting results.
Fig. 4 is a schematic view of robot obstacle avoidance in the same direction.
Fig. 5 is a schematic view of obstacle avoidance where the trajectory intersects the robot (faster speed).
Fig. 6 is a schematic view of obstacle avoidance where the trajectory intersects the robot (slower speed).
Fig. 7 is a forward direction collision schematic.
Fig. 8 is a schematic view of a trajectory intersection collision.
Fig. 9 is a schematic diagram of a six-wheeled omnidirectional mobile robot kinematic model.
Fig. 10 is a schematic view of a collision state.
Fig. 11 is a schematic view of a potential collision state.
Fig. 12 is a schematic view of a safe state.
Fig. 13 is a view of obstacle avoidance effect.
Detailed Description
According to the six-wheel independent driving independent steering robot obstacle avoidance path solving method and the obstacle avoidance strategy selection based on fuzzy decision, the robot obstacle avoidance is combined with track tracking control, the obstacle avoidance precision is improved through tracking feedback of the robot position, and the obstacle avoidance problem in static obstacle and dynamic obstacle environments is solved. It should be explained that, in the present invention, the point cloud data obtained by the depth camera and the lidar sensor is processed and sent to the robot after the obstacle information is obtained, and the present invention is mainly aimed at the obstacle avoidance process after the obstacle information is obtained, and the detailed description about the obtaining of the obstacle information is omitted.
Besides, the invention also provides a new model of six-wheel independent driving robots and a new situation of three obstacle avoidance problems in the field of obstacle avoidance of mobile robots, can further explore more obstacle avoidance type research contents of the mobile robots based on the model, and has high popularization value.
The invention provides an obstacle avoidance method and an obstacle avoidance strategy for a six-wheel independent driving independent steering robot, wherein the specific flow is shown in figure 1 and specifically comprises the following steps.
Establishing six-wheel kinematic model
The robot needs to track the path after acquiring the obstacle avoidance path, and the track tracking process is to calculate an optimal control sequence according to a kinematic model of the robot. Let the robot be a rigid body with non-deformable wheels, the sideslip angle β=0rad. The MPC track tracking controller used in the invention carries out obstacle avoidance path tracking, and the direct control quantity is the speed of the middle wheel and the angle of the left front wheel, so the speed of the middle wheel and the angle of the left front wheel are used as the control quantity in the analysis of a kinematic model. Therefore, in the coordinate system shown in fig. 9, the kinematic model of the robot is as follows:
wherein, the liquid crystal display device comprises a liquid crystal display device,
representing the position and yaw angle of the robot center, [ v delta ] l ] T For controlling the input of the quantity, namely the speed of the robot and the angle of rotation of the left front wheel. Delta l 、δ r The left front wheel corner and the right front wheel corner are respectively the same as the front wheels in terms of the left rear wheel corner and the right rear wheel corner, but are opposite in direction, and the description thereof is omitted. L is the distance between the front wheel and the middle wheel axle (also the distance between the middle wheel and the rear wheel axle), and D is the width of the robot. The invention does not simplify the width of the robot. According to the angular relationship, the relationship of the left wheel corner and the right wheel corner is as follows:
it can be deduced therefrom that the relationship between the turning radius and the front wheel deflection angle is:
the rear wheels of the six-wheel omnidirectional mobile robot adopt deflection directions with the same size and opposite directions as the front wheel steering angles, so the problem of rear wheel angles is not repeated here.
The robot is subject to incomplete constraints, i.e. the wheels of the robot do not slip, only a pure rolling motion, so this constraint is described by the following equation:
as a wheeled mobile robot, a state space model is most suitable for model prediction. From the above, the state space expression based on the robot kinematic model is:
wherein, the liquid crystal display device comprises a liquid crystal display device,
the state quantity is the state quantity in the kinematic model, and represents the central position and yaw angle of the robot, and u= [ v delta ] l ] T Is the control quantity of the robot.
The model predictive control algorithm is a bottom algorithm of a robot track tracking controller, and error compensation is carried out through a sliding mode control algorithm after planning and tracking are carried out by model predictive control, so that the nonlinear model is simplified into a linear model for processing, discretizing and then predictive control is needed for improving control efficiency.
The model is subjected to linearization treatment by a Taylor first-order expansion method, and in the motion process of the robot, the target state expected to be reached is a reference state, so that the Taylor first-order expansion is performed at the reference state:
wherein, the liquid crystal display device comprises a liquid crystal display device,
X r in order to take the examination state of machine, u r For the control quantity reference state, the A and the knife are jacobian matrixes, which are the partial derivatives of the function f on the state quantity and the control quantity respectively, and thus, the nonlinear kinematic model is converted into the linear kinematic model.
The forward Euler method is used for discretizing the formula within the sampling time T to obtain the following steps:
wherein I is an identity matrix. The above results are obtained by linearizing and discretizing the robot after kinematic modeling, and are input as a model of the MPC.
Secondly, obstacle track prediction
The prediction of the obstacle track is only used for judging the states of the robot and the obstacle, and is not used for accurately predicting the obstacle track.
In the collision prevention process of the robot, an obstacle track prediction link is required to be added. If the robot and the obstacle cannot collide with each other in the window in a certain time in the future, the obstacle avoidance is not needed; on the contrary, an appropriate obstacle avoidance strategy needs to be selected according to the position states of the robot and the obstacle. Therefore, the position conditions of the robot and the obstacle are classified into three for collision prediction, potential collision state, and safety state, as shown in fig. 10, 11, and 12. The track of the robot is used as a reference track to be input, and the track condition of the obstacle is mainly judged. The motion prediction calculation formula of the obstacle can be obtained by analyzing the Lagrangian basic polynomial and the truncated error formula thereof, and is as follows:
Wherein [ x ] kt ,y kt ]Represents k t Predicted obstacle position after each sampling instant, [ x ] k ,y k ]、[x k+1 ,y k+1 ]、[x k+2 ,y k+2 ]Respectively represent already t k Time t k+1 Time sum t k+2 The position of the obstacle observed at the moment. The motion trail of the future obstacle can be predicted according to the position of the obstacle at a known moment by using an obstacle motion prediction formula, so that collision judgment is carried out, and an obstacle avoidance strategy is selected。
The states of the robot and the obstacle are judged through the prediction of the obstacle track, and the distance between the robot and the obstacle is mainly used for judging after a certain time. If the distance between the robot and the obstacle is smaller than the safety margin distance, setting the state as a collision state, and selecting a proper obstacle avoidance strategy according to the collision type; if the distance between the robot and the obstacle is slightly larger than the safety margin distance, setting the robot to be in a potential collision state, and continuing to predict the obstacle track and analyze the state at the next moment; if the distance between the robot and the obstacle is far greater than the safety margin, the robot is set to be in a safe state, and the robot cannot be considered to collide with the obstacle.
Third, judging the threshold function
Different threshold functions are designed for different obstacle avoidance types to judge whether collision can occur, a safe distance threshold function is designed for the first case (collision state) to judge, and collision threshold functions are designed for the second case and the third case (potential collision state and safe state) to judge.
Safety distance threshold function:
aiming at the situation that the movement direction of the obstacle is the same as that of the robot, a safety distance threshold function is designed to judge whether a collision-free safety path needs to be planned, the robot returns to the original track to continue track tracking after the obstacle avoidance action is completed, and as shown in fig. 7, the distance threshold function is defined as follows:
wherein v is c And v obs A represents the movement speed of the robot and the movement speed of the obstacle respectively c And a obs Respectively representing the maximum acceleration of the robot and the maximum acceleration of the obstacle, t c D represents the time when the robot controller sends out instructions to the bottom motion module to react s1 Representing a minimum safety distance and a robot safety inflation margin.
Collision threshold function:
and aiming at the condition that the movement direction of the obstacle is intersected with the movement direction of the robot, designing a collision threshold function to judge whether a collision-free safety path needs to be planned. The determination is mainly made by the current moment and predicting the distance between the robot and the mass center of the obstacle for t sampling time steps in the future (the t-th moment is the moment when the obstacle moves to the intersection point with the robot track), as shown in fig. 8. The collision threshold function is defined as follows:
D th2 ＝min(d i (k)，d i (k+1)，…d i (k+t))+d s2
wherein d i (k) Represents the distance between the robot and the mass center of the obstacle at the moment k, d s2 Representing the robot safety inflation margin.
Solving obstacle avoidance path
The process of solving the obstacle avoidance path is that after the sensor detects an obstacle, obstacle information is sent to an obstacle avoidance controller and an obstacle avoidance planner, then an optimal path point is solved through constraints (distance constraints, control quantity constraints and the like) of an actual environment, an optimal path is obtained after track smoothing and polynomial fitting, the optimal path is discretized into coordinate points and then is output, an obstacle avoidance strategy is selected after the coordinate points are output (specific strategies are described below), obstacle avoidance is performed according to the strategies, and the whole obstacle avoidance process is finished. The whole obstacle avoidance process is shown in figure 1.
In the process of planning the obstacle avoidance path, certain expansion processing is required according to the size of the robot and the size of the obstacle, the obstacle is expanded according to the obstacle, the necessary safety margin of the robot and the information of the obstacle, and the expansion range of the robot is added to the part because of the safety margin, so the robot does not need to perform the expansion processing. The obstacle is subjected to the dividing process in consideration of the fact that there may be a large obstacle and the robot may pass through the middle of the obstacle, as shown in fig. 2.
In solving the obstacle avoidance path, an obstacle avoidance penalty function needs to be designed. The obstacle avoidance penalty function is mainly used for solving and restraining the obstacle avoidance function. The basic idea is to adjust the size and distance of the function by the distance deviation between the robot mass center coordinates and the barrier dividing pointsThe closer the distance, the larger the function value. Design obstacle avoidance penalty function P obs The following are provided:
wherein p is obs Is the barrier avoidance penalty function weight coefficient, (x) i ，y i ) Is the coordinates of the dividing points of the obstacle, and E is an extremely small positive number, so that the situation that 0 occurs in the denominator is avoided.
Besides the obstacle avoidance penalty function, in the solution of the obstacle avoidance function, the cost of the robot deviating from the reference track is designed as a constraint condition of the solution, and the cost function of the deviation reference track is designed as follows:
wherein x is min ，x max ，y min ，y max Is the global map coordinate range, x r ，y r Is the reference track coordinate, x a ，y a Is a coordinate point of the obstacle avoidance path.
Solving the obstacle avoidance path directly obtains the coordinates of the optimal path point, so that definition is that:
ξ＝[x a y a ] T ，
wherein x is a y a Is a coordinate point of the obstacle avoidance path.
In order to improve the accuracy of obstacle avoidance, a distance penalty function D is defined obs Adding soft constraint into obstacle avoidance function
Wherein K is a very large positive integer, and the method mainly aims at the situation that a dynamic obstacle suddenly approaches the robot in the moving process of the robot. When the distance between the robot and the obstacle is relatively short, the soft constraint item is relatively large, and the weight influence is large; when the distance between the robot and the obstacle is far, the soft constraint term is smaller, and the weight influence is small. The distance penalty function is:
wherein x is obs ，y obs For the heart coordinate of the obstacle, x rob ，y rob Is the robot mass heart coordinate d obs Weighting coefficients are distance penalty functions. The main control objective of the obstacle avoidance planner is to avoid obstacles and minimize deviation from a reference path. Therefore, according to the punishment items and constraints, the obstacle avoidance planner based on the mathematical optimization method is designed as follows:
wherein P is a weight matrix.
The control quantity of MPC is mainly used for restraining the obstacle avoidance path in the obstacle avoidance planner, and the obstacle avoidance path obtained by the planner is ensured to conform to the restraint of the robot kinematic model. The result of solving by the obstacle avoidance planner is the obstacle avoidance path of the robot.
In the constraint conditions of the obstacle avoidance function, the first term is the hard constraint of the robot control quantity, and the combination of the constraint conditions is the key of solving. And according to the deviation condition of the obstacle avoidance penalty function and the reference path, solving an optimal path in the range. The sampling time of the obstacle avoidance planner is the same as that of the track tracking controller, and a track tracking strategy with short prediction step length is adopted after the obstacle avoidance path is planned, so that the model prediction track tracking controller with better precision completely meets the requirement of path re-planning, and the control requirement of the robot can be met in real-time aspect.
The obstacle avoidance is mainly to determine an obstacle avoidance path region around an obstacle, and then find coordinates of a specific solving path in the region, namely the obstacle avoidance path.
In the obstacle avoidance controller algorithm, the obtained optimal solving sequence is the minimum deviation between the distance between the finite time domain and the reference point, and the obtained path points are given in the form of discrete points. As the condition of the obstacle and the condition of the reference trajectory change, the number and location of the planned local trajectory reference points also change. If the track is directly input into the track tracking controller, some redundant repeated track points increase the operation load of the track tracking controller, and the track tracking controller has different distribution densities of different reference points, so that the track tracking controller can hardly complete the track tracking task according to the discrete reference points.
By combining the situations, the obstacle avoidance path planned by the obstacle avoidance controller is required to be processed, so that the planner and the track tracking controller can be successfully connected. The obstacle avoidance path solved by the obstacle avoidance planner is mainly in the form of discrete points, and the requirements of smoothness and the like of the path are not considered, so that the smooth track points are mainly processed through curve fitting, and the curve fitting can generate track points which accord with the robot constraint (such as the robot position requirement is continuous, the yaw angle requirement is a first-order continuous curve, the acceleration constraint requirement is a second-order continuous curve and the like).
In the aspect of curve fitting, selecting n times of polynomial fitting as a fitting curve, and fitting based on the least square method principle, wherein the fitting function is as follows:
X＝a 1 x n +a 2 x n-1 +a 3 x n-2 +…+a n x+a n+1
Y＝b 1 y n +b 2 y n-1 +b 3 y n-2 +…+b n y+b n+1
wherein a= [ a ] 1 ，a 2 …a n+1 ]，b＝[b 1 ，b 2 …b n+1 ]The polynomial X and Y coefficients, respectively. Here, n=4 is chosen, i.e. the order of the fit is 4 times, and X and Y are the curve coordinates after the fit. Fig. 3 is a segmented fitting result of a re-planned track portion, and fig. 3 is a discrete track point after a track re-planning algorithm in the obstacle avoidance planner is smoothed by a curve, where the curve is a track curve after fitting.
Fifth, selection of obstacle avoidance strategy
The method mainly analyzes obstacle avoidance strategy selection of robots facing different obstacle avoidance conditions.
The obstacle avoidance problem disclosed by the invention is mainly used for researching the planning and tracking of obstacle avoidance tracks when the six-wheel omnidirectional mobile robot presents obstacles in the track tracking process. Therefore, the robot needs to select different obstacle avoidance tracks and strategies for different types of obstacles.
While planning the track, a certain control strategy needs to be designed to ensure that the robot can better complete the obstacle avoidance behavior. For a static obstacle, the static obstacle can be regarded as the condition that the speed of the dynamic obstacle is 0, and the obstacle avoidance track is planned and then tracked. For dynamic obstacles, three situations can be distinguished:
(1) In the first case, the obstacle movement direction is the same as the robot trajectory movement direction
As shown in FIG. 4, the planning of the obstacle avoidance track at this time mainly makes the robot choose the optimal track from the left side or the right side of the obstacle to overrun, and returns to the original track to track after overrun. And after the obstacle avoidance track is re-planned, outputting the new track points to a track tracking controller, and calculating the control quantity of the robot according to the MPC algorithm to track. However, when the position of the robot mass center and the position of the obstacle mass center are positioned on the same horizontal line, speed feedback proportion control is added into the speed item, so that the robot performs obvious overtaking action and returns to the original track as soon as possible. Meanwhile, differential control is added to prevent overshoot.
Wherein v is mpc The speed, k of the robot control quantity calculated by the MPC track tracking controller p1 And k d1 Is a speed feedback proportional control and differential control coefficient, e 11 Is the deviation value t of the obstacle avoidance track of the robot from the original track k Is the current moment of the robot movement, t d1max Is the robot moves to d shown in figure 4 1max Position moment, t d1min Is d in FIG. 4 1min The position moment shown.
(2) In the second case, the obstacle movement direction intersects with the robot movement direction, and the obstacle movement speed is high
As shown in fig. 5, the main difficulty in this case is to determine whether the robot tracking trajectory and the obstacle movement trajectory collide at the intersection point, and the main basis of the determination is related to the movement speeds of the robot and the obstacle. If the moving speed of the obstacle is relatively high, the robot is properly decelerated before reaching the collision point, then the robot passes through the rear of the obstacle after the obstacle passes through the intersection point, and then returns to the original track for tracking. The distance between the robot and the obstacle is small and then large, the speed needs to be reduced when the distance is small, the speed feedback proportional control is added on the basis of the MPC control quantity, and the MPC control is restored when the distance is large. Meanwhile, differential control is added to prevent overshoot when tracking the original track.
Wherein v is mpc The speed, k of the robot control quantity calculated by MPC p2 And k d2 Is a speed feedback proportional control and differential control coefficient, e 21 Is the deviation of the obstacle avoidance track of the robot from the original track, e 2 Is the distance d between the position of the robot mass center and the position of the obstacle mass center dis2 Is the absolute distance between the mass center of the robot and the mass center of the obstacle in the x direction, d 2min And d 2max The distance threshold value of the robot mass center and the obstacle is respectively.
(3) In the third case, the obstacle movement direction intersects with the robot movement direction, and the obstacle movement speed is slow
As shown in fig. 6, the robot accelerates properly before reaching the collision point and passes in front of the obstacle when the obstacle does not pass the intersection point, and then returns to the original trajectory to track. The distance between the robot and the obstacle is small and then large, the speed needs to be increased when the distance is small, the speed feedback proportional control is added on the basis of the MPC control quantity, the obstacle avoidance is accelerated, and the MPC control is restored when the distance is large. Meanwhile, differential control is added to prevent overshoot when tracking the original track.
Wherein v is mpc The speed, k of the robot control quantity calculated by MPC p3 And k d3 Is a speed feedback proportional control and differential control coefficient, e 31 Is the deviation of the obstacle avoidance track of the robot from the original track, e 3 Is the distance d between the position of the robot mass center and the position of the obstacle mass center dis3 Is the absolute distance between the mass center of the robot and the mass center of the obstacle in the x direction, d 3min And d 3max The distance threshold value of the robot mass center and the obstacle is respectively.
The main reason that the obstacle avoidance strategy is added with the speed feedback proportional differential control is that the obstacle avoidance track is different from the MPC track tracking algorithm of the original track, so that differential control is needed to be added at the junction for speed transition, and the tracking accuracy is ensured. Because the obstacle avoidance track is shorter, the strategy of short prediction step length and weight matrix is adopted in the MPC track tracking controller, and compared with the tracking of the common track, the calculation efficiency and the obstacle avoidance timeliness are improved.
The robot obstacle avoidance strategy is mainly selected by predicting an obstacle track and a robot track, predicting whether collision occurs or not, and then performing strategy selection. For static obstacles, the situation that the movement speed and the acceleration are 0 can be considered, and when the safety distance threshold is exceeded, obstacle avoidance track planning is conducted.
For a dynamic obstacle, if the moving track of the obstacle is a straight line, directly judging whether collision occurs after a certain time or not according to the calculation result of the threshold function, and then solving an optimal path by utilizing constraint; if the moving track of the obstacle is a curve, predicting the track after a certain time according to the differential function of the displacement of the obstacle to the time, judging the movement condition of the robot under the time according to the movement time when the robot intersects with the track of the robot, judging whether collision occurs or not by utilizing the centroid distance of the robot and the obstacle, and then solving the optimal path.
In robot obstacle avoidance control, the present invention provides a rigid and systematic approach that takes into account the safety constraints of the robot and is used in a closed loop with a feedback controller.
And adopting an asymmetric model in a fuzzy decision in the selection of the robot obstacle avoidance decision, taking the obstacle avoidance condition of the robot as a constraint condition, and taking the selection of the robot obstacle avoidance strategy as an objective function. In an asymmetric model, taking the acceptance constraint as a prerequisite, the position of both the target and the constraint is not symmetric. Given the objective function f (X) on the argument X and the constraint condition fuzzy set D on X, the optimal solution M for maximizing f under the constraint D is a fuzzy subset on X, which has membership functions μm (X) equal to 0 when the set on the right end of the equation is the empty set, the given argument X is the set of all cases, the current situation encountered by the robot is the constraint D, and the objective function is the obstacle avoidance strategy of the robot.
The robot solves the objective function by taking the environment information and the obstacle information when encountering the obstacle as constraint conditions, and selects a proper obstacle avoidance strategy.
Sixth, robot keeps away barrier
The effect of obstacle avoidance is shown in fig. 13 after the planning of the obstacle avoidance path and the selection of the obstacle avoidance strategy are performed. It can be seen that the robot well solves the obstacle avoidance problem, and the obstacle avoidance is realized under the condition of encountering an obstacle.
Claims (8)
1. An obstacle avoidance method for a six-wheel independent driving independent steering robot is characterized by comprising the following steps of: solving an obstacle avoidance path and an obstacle avoidance strategy based on fuzzy decision; after the sensor detects the obstacle, the obstacle information is sent to the obstacle avoidance controller and the obstacle avoidance planner, then the optimal path point is solved through the constraint of the actual environment, the optimal path is obtained through track smoothing and polynomial fitting, the optimal path is output after being scattered into coordinate points, the obstacle avoidance strategy is selected after the coordinate points are output, the obstacle avoidance is carried out according to the strategy, and the whole obstacle avoidance process is finished.
2. The obstacle avoidance method for a six-wheeled independently driven independently steered robot of claim 1, characterized by: the process for solving the obstacle avoidance path is as follows:
in the process of planning an obstacle avoidance path, performing expansion processing according to the size of the robot and the size of the obstacle, and expanding the obstacle according to the obstacle, the necessary safety margin of the robot and the information of the obstacle, wherein the expansion range of the robot is added into the part because of the safety margin; taking the situation that the obstacle is large and the robot possibly passes through the middle of the obstacle into consideration, dividing the obstacle, and setting an obstacle dividing point at the outline of the obstacle to perform dividing;
In the solving process of the obstacle avoidance path, designing an obstacle avoidance penalty function and a cost function of a design deviation reference track; the obstacle avoidance penalty function is used for solving and restraining the obstacle avoidance function, the size of the function is adjusted through the distance deviation between the robot mass center coordinates and the obstacle dividing points, and the function value is larger as the distance is closer; taking a cost function of the robot deviated from the reference track as a constraint condition for solving;
solving the obstacle avoidance path directly obtains the coordinates of the optimal path point, so that definition is that:
ξ＝[x a y a ] T ，
wherein x is a y a The coordinate points are obstacle avoidance path coordinate points;
in order to improve the accuracy of obstacle avoidance, a distance penalty function D is defined obs In the obstacle avoidance functionAdding soft constraints
K is a very large positive integer, which is aimed at the situation that a dynamic obstacle suddenly approaches the robot in the moving process of the robot, when the distance between the robot and the obstacle is relatively short, the soft constraint item is relatively large, and the weight influence is large; when the distance between the robot and the obstacle is far, the soft constraint item is smaller, and the weight influence is small;
according to the punishment items and constraints, designing a barrier avoidance planner based on a mathematical optimization method, wherein the barrier avoidance planner is designed as follows:
wherein: p is a weight matrix;
the term is the control quantity of the MPC, and the action in the obstacle avoidance planner is to restrict the obstacle avoidance path, so that the obstacle avoidance path obtained by the planner is ensured to conform to the restriction of the robot kinematic model; the result xi solved by the obstacle avoidance planner is the obstacle avoidance path of the robot; p (P) obs To avoid barrier penalty function D obs Distance penalty function. In the constraint condition, a->
Ensuring that the result accords with a motion model of the robot for MPC control quantity, wherein gamma is a cost function of a deviation reference track;
and processing the obstacle avoidance path solved by the obstacle avoidance planner through curve fitting, wherein the curve fitting generates the smoothed track points into track points conforming to the constraint of the robot.
3. According to claimThe obstacle avoidance method of the six-wheel independent driving independent steering robot, which is characterized in that: the obstacle avoidance penalty function P obs The following are provided:
wherein p is obs Is the barrier avoidance penalty function weight coefficient, (x) i ,y i ) Is the coordinates of the dividing points of the obstacle, and E is an extremely small positive number, so that the situation that 0 occurs in the denominator is avoided.
4. The obstacle avoidance method for a six-wheeled independently driven independently steered robot of claim 2, characterized by: the cost function of the deviation reference track is as follows:
wherein x is min ,x max ,y min ,y max Is the global map coordinate range, x r ，y r Is the reference track coordinate, x a ，y a Is a coordinate point of the obstacle avoidance path.
5. The obstacle avoidance method for a six-wheeled independently driven independently steered robot of claim 2, characterized by: the distance penalty function is:
wherein x is obs ，y obs For the heart coordinate of the obstacle, x rob ，y rob Is the robot mass heart coordinate d obs Weighting coefficients are distance penalty functions.
6. The obstacle avoidance method for a six-wheeled independently driven independently steered robot of claim 2, characterized by: and the curve fitting is carried out by selecting n times of polynomial fitting as a fitting curve and carrying out fitting based on the least square method principle, wherein the fitting function is as follows:
X＝a 1 x n +a 2 x n-1 +a 3 x n-2 +…+a n x+a n+1 ，
Y＝b 1 y n +b 2 y n-1 +b 3 y n-2 +…+b n y+b n+1 ，
wherein a= [ a ] 1 ,a 2 …a n+1 ],b＝[b 1 ,b 2 …b n+1 ]Polynomial X and Y coefficients, respectively; n=4 is chosen, i.e. the order of the fit is 4 times, and X and Y are the coordinates of the curve after the fit.
7. The obstacle avoidance method for a six-wheeled independently driven independently steered robot of claim 1, characterized by: the obstacle avoidance strategy based on fuzzy decision is as follows:
the robot obstacle avoidance strategy is selected by predicting an obstacle track and a robot track, predicting whether collision occurs or not, and then performing strategy selection; when the static obstacle is regarded as the condition that the motion speed and the acceleration are 0 and the safety distance threshold value is exceeded, the obstacle avoidance track is planned, and track tracking is performed after the obstacle avoidance track is planned;
for a dynamic obstacle, if the moving track of the obstacle is a straight line, directly judging whether collision occurs after a certain time or not according to the calculation result of a threshold function, and then solving an optimal path by utilizing constraint; if the moving track of the obstacle is a curve, predicting the track after a certain time according to the differential function of the displacement of the obstacle to the time, judging the movement condition of the robot under the time according to the movement time when the robot intersects with the track of the robot, judging whether collision occurs or not by utilizing the centroid distance of the robot and the obstacle, and then solving the optimal path.
An asymmetric model in a fuzzy decision is adopted in the selection of the robot obstacle avoidance decision, the obstacle avoidance condition of the robot is taken as a constraint condition, and the selection of the robot obstacle avoidance strategy is taken as an objective function; in an asymmetric model, taking acceptance constraints as preconditions, the position of both the target and the constraints is not symmetric; giving an objective function f (X) on a domain X and a constraint condition fuzzy set D on X, wherein an optimal solution m for maximizing f under the constraint D is a fuzzy subset on X, the fuzzy subset has a membership function, when the set on the right end of an equation is an empty set, mum (X) is equal to 0, the given domain X is a set of all conditions, the current condition encountered by the robot is the constraint D, and the objective function is an obstacle avoidance strategy of the robot;
the robot solves the objective function by taking the environment information and the obstacle information when encountering the obstacle as constraint conditions, and selects a proper obstacle avoidance strategy.
8. The obstacle avoidance method for a six-wheeled independently driven independently steered robot of claim 7, wherein: the dynamic barrier is specifically divided into the following three cases:
(1) In the first case, the movement direction of the obstacle is the same as the movement direction of the robot track;
At the moment, the planning of the obstacle avoidance track is to enable the robot to select an optimal track from the left side or the right side of the obstacle to overrun, and return to the original track for tracking after overrun; after the obstacle avoidance track is re-planned, outputting new track points to a track tracking controller, and calculating the control quantity of the robot according to an MPC algorithm to track; when the position of the robot mass center and the position of the obstacle mass center are positioned on the same horizontal line, adding speed feedback proportional control into a speed item to enable the robot to perform obvious overtaking operation and return to an original track as soon as possible, and adding differential control to prevent overshoot;
wherein v is mpc The speed, k of the robot control quantity calculated by the MPC track tracking controller p1 And k d1 Is a speed feedback proportional control and differential control coefficient, e 11 Is the deviation value t of the obstacle avoidance track of the robot from the original track k Is the current moment of the robot movement, t d1max Is the robot moves to d 1max Position moment, t d1min Is d 1min The time of the position is shown;
(2) In the second case, the movement direction of the obstacle intersects with the movement direction of the robot, and the movement speed of the obstacle is higher;
the situation is that whether the robot tracking track and the obstacle moving track collide at the intersection point or not is judged, and the judgment basis is related to the moving speed of the robot and the obstacle; if the moving speed of the obstacle is relatively high, the robot is properly decelerated before reaching the collision point, then the robot overruns from the rear of the obstacle after the obstacle passes over the intersection point, and then returns to the original track for tracking; the distance between the robot and the obstacle is small and then large, the speed needs to be reduced when the distance is small, the speed feedback proportional control is added on the basis of the MPC control quantity, the MPC control is restored when the distance is large, and meanwhile the differential control is added to prevent overshoot when the original track is tracked;
Wherein v is mpc The speed, k of the robot control quantity calculated by MPC p2 And k d2 Is a speed feedback proportional control and differential control coefficient, e 21 Is the deviation of the obstacle avoidance track of the robot from the original track, e 2 Is the distance d between the position of the robot mass center and the position of the obstacle mass center dis2 Is the absolute distance between the mass center of the robot and the mass center of the obstacle in the x direction, d 2min And d 2max The distance threshold value of the mass center of the robot and the obstacle is respectively;
(3) In the third case, the movement direction of the obstacle intersects with the movement direction of the robot, and the movement speed of the obstacle is slower;
the robot is accelerated properly before reaching the collision point, and overruns from the front of the obstacle when the obstacle does not cross the intersection point, and then returns to the original track for tracking; the distance between the robot and the obstacle is small and then large, the speed needs to be increased when the distance is small, the speed feedback proportional control is added on the basis of the MPC control quantity, the obstacle avoidance is accelerated, the MPC control is restored when the distance is large, and meanwhile the differential control is added to prevent overshoot when the original track is tracked;
wherein v is mpc The speed, k of the robot control quantity calculated by MPC p3 And k d3 Is a speed feedback proportional control and differential control coefficient, e 31 Is the deviation of the obstacle avoidance track of the robot from the original track, e 3 Is the distance d between the position of the robot mass center and the position of the obstacle mass center dis3 Is the absolute distance between the mass center of the robot and the mass center of the obstacle in the x direction, d 3min And d 3max The distance threshold value of the robot mass center and the obstacle is respectively.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310103879.1A CN116225004A (en) | 2023-02-13 | 2023-02-13 | Obstacle avoidance method for six-wheel independent driving independent steering robot |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310103879.1A CN116225004A (en) | 2023-02-13 | 2023-02-13 | Obstacle avoidance method for six-wheel independent driving independent steering robot |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116225004A true CN116225004A (en) | 2023-06-06 |
Family
ID=86576222
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310103879.1A Pending CN116225004A (en) | 2023-02-13 | 2023-02-13 | Obstacle avoidance method for six-wheel independent driving independent steering robot |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116225004A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117341683A (en) * | 2023-12-04 | 2024-01-05 | 苏州观瑞汽车技术有限公司 | Vehicle dynamic track fitting obstacle avoidance method and system based on multi-target recognition |
-
2023
- 2023-02-13 CN CN202310103879.1A patent/CN116225004A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117341683A (en) * | 2023-12-04 | 2024-01-05 | 苏州观瑞汽车技术有限公司 | Vehicle dynamic track fitting obstacle avoidance method and system based on multi-target recognition |
| CN117341683B (en) * | 2023-12-04 | 2024-04-23 | 苏州观瑞汽车技术有限公司 | Vehicle dynamic track fitting obstacle avoidance method and system based on multi-target recognition |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110377039B (en) | Vehicle obstacle avoidance track planning and tracking control method | |
| Shan et al. | A reinforcement learning-based adaptive path tracking approach for autonomous driving | |
| Maalouf et al. | A higher level path tracking controller for a four-wheel differentially steered mobile robot | |
| Peng et al. | Leader–follower formation control of nonholonomic mobile robots based on a bioinspired neurodynamic based approach | |
| CN109799814B (en) | Design method of track unmanned vehicle track tracking controller based on kinematics model | |
| CN108227491B (en) | Intelligent vehicle track tracking control method based on sliding mode neural network | |
| Awad et al. | Model predictive control with fuzzy logic switching for path tracking of autonomous vehicles | |
| Shi et al. | Planning the trajectory of an autonomous wheel loader and tracking its trajectory via adaptive model predictive control | |
| Ding et al. | Trajectory tracking of redundantly actuated mobile robot by MPC velocity control under steering strategy constraint | |
| Ruslan et al. | Modelling and control strategies in path tracking control for autonomous tracked vehicles: A review of state of the art and challenges | |
| CN116225004A (en) | Obstacle avoidance method for six-wheel independent driving independent steering robot | |
| Hegedüs et al. | Motion planning for highly automated road vehicles with a hybrid approach using nonlinear optimization and artificial neural networks | |
| CN112937571A (en) | Intelligent automobile track tracking control method and system | |
| Qie et al. | An improved model predictive control-based trajectory planning method for automated driving vehicles under uncertainty environments | |
| Rasib et al. | Are self-driving vehicles ready to launch? An insight into steering control in autonomous self-driving vehicles | |
| Lee et al. | Learning terrain-aware kinodynamic model for autonomous off-road rally driving with model predictive path integral control | |
| Akiba et al. | Optimal tracking control of two-wheeled mobile robots based on model predictive control | |
| Jiang et al. | Learning based predictive error estimation and compensator design for autonomous vehicle path tracking | |
| Fnadi et al. | Local obstacle-skirting path planning for a fast bi-steerable rover using bézier curves | |
| Zheng et al. | Model predictive control for intelligent vehicle lane change | |
| Kim et al. | Vehicle path tracking control using pure pursuit with MPC-based look-ahead distance optimization | |
| Reda et al. | Model-based control strategy for autonomous vehicle path tracking task | |
| Pérez-Morales et al. | Information-Theoretic Sensor-Based Predictive Control for Autonomous Vehicle Navigation: A Proof of Concept | |
| Kim et al. | Hierarchical Motion Planning and Control Algorithm of Autonomous Racing Vehicles for Overtaking Maneuvers | |
| Simotwo et al. | Control strategies to steer and drive an autonomous 4WS4WD ground vehicle: A review of MPC approaches |
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 |