Local motion planning method for high-agility wheeled vehicles in response to sudden aerial risks

Abstract

The application belongs to the technical field of motion planning, and particularly relates to a local motion planning method for a high-agility wheeled vehicle coping with sudden air risks, which is based on composite risk assessment of an artificial potential field, maps risk positions and speed attributes to a spatial potential field model and a speed potential field model, completes initial screening of the risks, analyzes collision points according to the initial screening result, completes fine screening of the risks, and generates a risk assessment value in combination with risk categories; performs local motion sampling according to vehicle states and performance constraint information, and generates an expected candidate trajectory set; calculates a performance function of each candidate trajectory based on the risk assessment value, evaluates an optimal trajectory meeting risk avoidance and vehicle driving stability, and selects planning control instructions of the optimal trajectory in real time according to a rolling optimization theory, and sends the planning control instructions to a chassis controller for execution, so that air risk avoidance is dynamically realized.

Description

Technical Field

[0001] This invention belongs to the field of motion planning technology, specifically relating to a highly agile local motion planning method for wheeled vehicles to cope with sudden aerial risks. Background Technology

[0002] In urban construction site scenarios, vehicles face significant safety challenges due to the unpredictable risk of falling objects from the air. Under such conditions, the extremely short window for handling risk stress cannot be sustained by human reaction alone. Therefore, vehicles are required to possess strong intelligent maneuverability, autonomously taking over the vehicle control system to agilely and proactively maneuver and avoid sudden aerial risks, thereby ensuring their safety. Most existing path planning methods based on search, sampling, and optimization are primarily focused on ground risks, neglecting the impact of vehicle voxel height on aerial risk assessment in a three-dimensional environment. These methods are therefore unsuitable for high-speed aerial risk assessment and evasive maneuver planning. Summary of the Invention

[0003] (a) Technical problems to be solved

[0004] The technical problem to be solved by this invention is: to provide a highly agile local motion planning method for wheeled vehicles to cope with sudden aerial risks, so as to realize the active maneuvering and evasive movement of the vehicle, develop the intelligent stress response characteristics of the vehicle, and improve its survivability.

[0005] (II) Technical Solution

[0006] To address the aforementioned technical problems, this invention provides a highly agile local motion planning method for wheeled vehicles to cope with sudden aerial risks. The local motion planning method includes the following steps:

[0007] Step 1: Initial Risk Screening;

[0008] After detecting incoming aerial risks, the system quickly predicts risk information and vehicle trajectories. Based on a composite risk assessment using an artificial potential energy field, the system maps the risk location and velocity attributes to a spatial potential field model and a velocity potential field model to complete the initial risk screening.

[0009] Step 2: Risk screening section;

[0010] Based on the initial screening results, collision point analysis is performed to screen out the predicted aerial risk trajectories that will hit the vehicle, thus completing the risk fine screening. The risk threat coefficient is generated by combining the risk category coefficient.

[0011] Step 3: Candidate trajectory sampling;

[0012] Based on vehicle status and performance constraint information, local expected trajectory sampling is performed to generate a set of feasible candidate trajectories for vehicles within the prediction domain.

[0013] Step 4: Optimal trajectory evaluation;

[0014] The performance evaluation function value of each candidate trajectory is calculated based on the risk threat coefficient, and the optimal trajectory that satisfies risk avoidance and vehicle driving stability is evaluated.

[0015] Step 5: Generate the optimal planning trajectory;

[0016] Based on the rolling optimization theory, the optimal trajectory planning instruction is selected in real time and sent to the chassis controller for execution, thereby dynamically realizing aerial risk avoidance.

[0017] Through the above steps, the highly agile wheeled vehicle local motion planning method for responding to sudden aerial risks can automatically assess aerial risks and calculate risk avoidance trajectories in real time.

[0018] In the initial risk screening part of step 1, the artificial potential energy field adopts a coupled potential energy field model based on the spatial potential field model and the velocity potential field model.

[0019] The spatial potential field model takes the coordinates of aerial risks and vehicle coordinates as the research objects. It analyzes whether the danger threshold is reached at a future time. If it is reached, it is determined that the aerial risk will pose a threat to the vehicle at a future time, and the aerial risk is the threat target.

[0020] The velocity potential field model takes the air risk and the vehicle's current velocity vector as the research object; it determines whether the angle between the current composite velocity vector and the risk pointing vector is within the expansion angle, and if so, it is determined to be a risk target.

[0021] In the risk screening part of step 2, risk category coefficients are obtained by mapping and evaluating the predicted aerial risk trajectory based on the identified target category.

[0022] In the aerial risk and threat determination part of steps 1 and 2, the aerial risk prediction trajectory point sequence and the vehicle prediction trajectory point sequence are input. After Euclidean distance coarse screening, vehicle coordinate system transformation and coordinate threshold fine screening, the aerial threat and risk determination at the current moment is completed based on the spatial field of view model and the velocity field of view model.

[0023] If it is determined that an aerial threat will hit the vehicle, output the remaining hit time, the coordinates of the hit point, and the velocity vector parameters of the hit point to subsequent steps;

[0024] If it is determined that the vehicle will not be hit, the time of impact and the predicted landing point coordinates will be output to subsequent steps.

[0025] If the incoming aerial object has not yet landed, the process proceeds to the next sampling time to reassess the aerial threat under the new conditions.

[0026] In step 2, the aerial risk threat assessment section, if it is determined that an aerial threat will hit the vehicle, calculates the risk threat coefficient based on the risk category coefficient corresponding to the identified target category and the remaining hit time, and outputs it to the subsequent steps for optimal trajectory assessment.

[0027] In the local expected trajectory sampling process of step 3, trajectory points are sampled based on the vehicle's own performance. The sampling constraints include vehicle performance constraints, vehicle geometric and mechanical constraints, vehicle sideslip dynamics constraints, and vehicle obstacle avoidance safety constraints.

[0028] In the optimal trajectory evaluation process in step 4, the performance evaluation function includes a target orientation evaluation term, a velocity evaluation term, and an obstacle distance evaluation term that take into account the target type.

[0029] The obstacle distance evaluation item is related to the risk threat coefficient.

[0030] In step 5, during the selection of the optimal trajectory planning instruction, within each sampling interval, the top three control instructions of the optimal trajectory with the highest performance evaluation function value are selected and sent to the chassis controller for platform maneuver control.

[0031] The method described above enables emergency trajectory planning in response to sudden aerial risks, improving the agility, intelligence, and safety of vehicle maneuverability, and enhancing the vehicle's survivability.

[0032] (III) Beneficial Effects

[0033] Compared with existing technologies, this invention provides a highly agile local motion planning method for wheeled vehicles to cope with sudden aerial risks. It introduces spatial potential field and velocity potential field to determine aerial risks, and can take into account the difference in impact point caused by vehicle elevation. Compared with two-dimensional ground obstacle avoidance planning strategies, this invention can handle scenarios where there is an aerial risk that will hit the vehicle with an angled flight trajectory. The judgment scheme is more reasonable and more suitable for mission scenarios with aerial risks. By introducing risk coefficient assessment based on risk category identification, the avoidance trajectory is automatically adjusted according to the target risk level, improving the safety of the planned trajectory. The planned local motion trajectory can simultaneously meet the vehicle's extreme performance constraints, vehicle's geometric and mechanical constraints, vehicle's sideslip dynamics constraints, and vehicle's obstacle avoidance safety constraints.

[0034] Through the above embodiments, the high-agility wheeled vehicle local motion planning method for responding to sudden aerial risks can realize stress trajectory planning for sudden aerial risks, improve the agility, intelligence and safety of vehicle maneuverability, and enhance the vehicle's survivability. Attached Figure Description

[0035] Figure 1This is a schematic diagram of the local planning strategy for dealing with aerial risks according to the present invention.

[0036] Figure 2 This invention provides a flowchart for calculating the local motion planning trajectory of a highly agile wheeled vehicle in response to sudden aerial risks. Detailed Implementation

[0037] To make the objectives, contents, and advantages of the present invention clearer, the specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples.

[0038] To address the aforementioned technical problems, this invention provides a highly agile local motion planning method for wheeled vehicles to cope with sudden aerial risks. The local motion planning method includes the following steps:

[0039] Step 1: Initial Risk Screening;

[0040] After detecting incoming aerial risks, the system quickly predicts risk information and vehicle trajectories. Based on a composite risk assessment using an artificial potential energy field, the system maps the risk location and velocity attributes to a spatial potential field model and a velocity potential field model to complete the initial risk screening.

[0041] Step 2: Risk screening section;

[0042] Based on the initial screening results, collision point analysis is performed to screen out the predicted aerial risk trajectories that will hit the vehicle, thus completing the risk fine screening. The risk threat coefficient is generated by combining the risk category coefficient.

[0043] Step 3: Candidate trajectory sampling;

[0044] Based on vehicle status and performance constraint information, local expected trajectory sampling is performed to generate a set of feasible candidate trajectories for vehicles within the prediction domain.

[0045] Step 4: Optimal trajectory evaluation;

[0046] The performance evaluation function value of each candidate trajectory is calculated based on the risk threat coefficient, and the optimal trajectory that satisfies risk avoidance and vehicle driving stability is evaluated.

[0047] Step 5: Generate the optimal planning trajectory;

[0048] Based on the rolling optimization theory, the optimal trajectory planning instruction is selected in real time and sent to the chassis controller for execution, thereby dynamically realizing aerial risk avoidance.

[0049] Through the above steps, the highly agile wheeled vehicle local motion planning method for responding to sudden aerial risks can automatically assess aerial risks and calculate risk avoidance trajectories in real time.

[0050] In the initial risk screening part of step 1, the artificial potential energy field adopts a coupled potential energy field model based on the spatial potential field model and the velocity potential field model.

[0051] The spatial potential field model takes the coordinates of aerial risks and vehicle coordinates as the research objects. It analyzes whether the danger threshold is reached at a future time. If it is reached, it is determined that the aerial risk will pose a threat to the vehicle at a future time, and the aerial risk is the threat target.

[0052] The velocity potential field model takes the air risk and the vehicle's current velocity vector as the research object; it determines whether the angle between the current composite velocity vector and the risk pointing vector is within the expansion angle, and if so, it is determined to be a risk target.

[0053] In the risk screening part of step 2, risk category coefficients are obtained by mapping and evaluating the predicted aerial risk trajectory based on the identified target category.

[0054] In the aerial risk and threat determination part of steps 1 and 2, the aerial risk prediction trajectory point sequence and the vehicle prediction trajectory point sequence are input. After Euclidean distance coarse screening, vehicle coordinate system transformation and coordinate threshold fine screening, the aerial threat and risk determination at the current moment is completed based on the spatial field of view model and the velocity field of view model.

[0055] If it is determined that an aerial threat will hit the vehicle, output the remaining hit time, the coordinates of the hit point, and the velocity vector parameters of the hit point to subsequent steps;

[0056] If it is determined that the vehicle will not be hit, the time of impact and the predicted landing point coordinates will be output to subsequent steps.

[0057] If the incoming aerial object has not yet landed, the process proceeds to the next sampling time to reassess the aerial threat under the new conditions.

[0058] In step 2, the aerial risk threat assessment section, if it is determined that an aerial threat will hit the vehicle, calculates the risk threat coefficient based on the risk category coefficient corresponding to the identified target category and the remaining hit time, and outputs it to the subsequent steps for optimal trajectory assessment.

[0059] In the local expected trajectory sampling process of step 3, trajectory points are sampled based on the vehicle's own performance. The sampling constraints include vehicle performance constraints, vehicle geometric and mechanical constraints, vehicle sideslip dynamics constraints, and vehicle obstacle avoidance safety constraints.

[0060] In the optimal trajectory evaluation process in step 4, the performance evaluation function includes a target orientation evaluation term, a velocity evaluation term, and an obstacle distance evaluation term that take into account the target type.

[0061] The obstacle distance evaluation item is related to the risk threat coefficient.

[0062] In step 5, during the selection of the optimal trajectory planning instruction, within each sampling interval, the top three control instructions of the optimal trajectory with the highest performance evaluation function value are selected and sent to the chassis controller for platform maneuver control.

[0063] The method described above enables emergency trajectory planning in response to sudden aerial risks, improving the agility, intelligence, and safety of vehicle maneuverability, and enhancing the vehicle's survivability.

[0064] Example 1

[0065] This embodiment provides a specific implementation method for local motion planning of highly agile wheeled vehicles to cope with sudden aerial risks.

[0066] I. Initial Risk Screening: After detecting the risk of an incoming air attack, the system quickly predicts the risk information and the vehicle's trajectory. Based on the composite risk assessment using an artificial potential energy field, the system maps the risk location and velocity attributes to the spatial potential field model and the velocity potential field model to complete the initial risk screening.

[0067] II. Risk Refinement Screening: Based on the initial screening results, collision point analysis is performed to identify high-risk trajectories that are expected to hit the vehicle, thus completing the risk refinement screening. Risk assessment values ​​are then generated by combining the risk categories.

[0068] III. Candidate Trajectory Sampling: Based on vehicle status and performance constraint information, local expected trajectory sampling is performed to generate a set of feasible candidate trajectories for vehicles within the prediction domain;

[0069] IV. Optimal Trajectory Evaluation: Based on the risk assessment value, calculate the performance function value of each candidate trajectory, and evaluate the optimal trajectory that satisfies both risk avoidance and vehicle driving stability.

[0070] V. Optimal trajectory generation: Based on the rolling optimization theory, the optimal trajectory planning instruction is selected in real time and sent to the chassis controller for execution, dynamically realizing aerial risk avoidance;

[0071] During implementation, in the initial screening phase, when an aerial target is detected in the vicinity and its trajectory sequence points are collected, the estimated time t and spatial coordinates p of the aerial target's arrival at the ground are calculated. mis,t And obtain the spatial coordinates p of the vehicle at that moment through trajectory prediction. t Risk assessment based on spatial potential field is performed by comparing the relationship between the distance d between the two and the threshold dr.

[0072]

[0073] Where index pA value of 1 indicates that the unit is at risk. p A value of 0 indicates that the unit has no risk;

[0074] During implementation, in the initial screening phase, when an aerial target is detected in the vicinity and its trajectory sequence points are collected, the data is then analyzed based on the current vehicle speed. And aerial threats, current movement speed is Calculate the resultant velocity vector Threat location vector And the angle θ corresponding to the radius of threat expansion dm. s Risk assessment based on velocity potential field:

[0075]

[0076] Where index v A value of 1 indicates that the unit is at risk. v A value of 0 indicates that the unit has no risk;

[0077] During implementation, in the fine screening section, the coordinates P of the trajectory point of the incoming object are... d (t) is converted to coordinates P in the vehicle coordinate system. d '(t); if P d If '(t) is not within the vehicle envelope threshold range, then the incoming object did not hit the vehicle at that moment, and the next moment is calculated; if P d If (t) is within the vehicle envelope threshold range, then it is determined that the incoming object has hit the vehicle at that moment, and this moment is denoted as t. k ;

[0078] Among them, the hit point P hit Set to P d '(t k-1 ) and P d '(t k The midpoint of P) hit =(P d '(t k-1 )+P d '(t k )) / 2; the hit time is T hit =(t k-1 +t k ) / 2; The direction vector of the incoming object (in vehicle coordinates) upon impact is

[0079] In the aerial risk threat assessment section, if it is determined that an aerial threat will hit the vehicle, then the risk category coefficient ξ corresponding to the target identification category is used. i and remaining hit time t re Calculate the risk factor iThis risk assessment value is output to the planning module for optimal trajectory evaluation.

[0080]

[0081] Where i is the risk number;

[0082] During the local expected trajectory sampling process, trajectory points are sampled based on the vehicle's own performance. The sampling constraint V mainly includes the vehicle's performance constraint V. m The vehicle's geometric and mechanical constraints V s Vehicle sideslip dynamics constraint V f And vehicle obstacle avoidance safety constraints V a :

[0083] V∈V m ∩V s ∩V f ∩V a

[0084]

[0085] V s ={(v,ω)|δ∈[δ min, δ max}

[0086] V f ={(v,ω)|v≤v r}

[0087]

[0088] Where v min v max ω min ω max δ min, δ max v represents the minimum / maximum values ​​of vehicle speed, angular velocity, and front wheel steering angle. c and ω c These are the vehicle's current speed and angular velocity, v. r The critical rollover speed of the vehicle. and These are the absolute values ​​of the acceleration and angular acceleration of the obstacle relative to the vehicle, respectively, and dist(v,ω) is the distance between the vehicle trajectory sampling point and the target risk.

[0089] The optimal trajectory evaluation, performance evaluation function G(v,ω), mainly includes three aspects: target orientation evaluation (heading(v,ω,type) considering target type), velocity evaluation (velocity(v,ω,type)), and obstacle distance evaluation (dist(v,ω,type)).

[0090] G(v,ω)=σ(α·heading(v,ω,type)+β·velocity(v,ω,type)+γ·dist(v,ω,type))

[0091] Where σ, α, ω min β and γ are the corresponding weight coefficients;

[0092] In dist(v,ω,type), the type attribute is determined by the risk factor. i Influence on decisions;

[0093] During the planning instruction selection process, within each control sampling interval, the top 3 control instructions based on the optimal trajectory with the highest target value of the performance evaluation function are selected and sent to the chassis controller for platform maneuver control.

[0094] Compared with existing technologies, this embodiment provides a highly agile local motion planning method for wheeled vehicles to cope with sudden aerial risks. It introduces spatial potential field and velocity potential field to determine aerial risks, and can take into account the difference in impact point caused by vehicle elevation. Compared with two-dimensional ground obstacle avoidance planning strategies, this embodiment can handle scenarios where there is an aerial risk that will hit the vehicle with an inclined flight trajectory. The judgment scheme is more reasonable and more suitable for mission scenarios with aerial risks. By introducing risk coefficient assessment based on risk category identification, the avoidance trajectory is automatically adjusted according to the target risk level, improving the safety of the planned trajectory. The planned local motion trajectory can simultaneously meet the vehicle's extreme performance constraints, vehicle's geometric and mechanical constraints, vehicle's sideslip dynamics constraints, and vehicle's obstacle avoidance safety constraints.

[0095] Through the above embodiments, the high-agility wheeled vehicle local motion planning method for responding to sudden aerial risks can realize stress trajectory planning for sudden aerial risks, improve the agility, intelligence and safety of vehicle maneuverability, and enhance the vehicle's survivability.

[0096] In summary, this invention belongs to the field of motion planning technology, specifically relating to a highly agile local motion planning method for wheeled vehicles to cope with sudden aerial risks. It solves the problem of low path feasibility caused by existing wheeled vehicle local motion planning algorithms neglecting the three-dimensional motion characteristics of aerial risks. This invention is based on a risk assessment algorithm under a composite artificial potential energy field, dynamically adjusting the vehicle's local motion planning control output to achieve a highly agile avoidance effect in response to sudden aerial risks. The process includes: a composite risk assessment based on an artificial potential energy field, mapping the risk location and velocity attributes to a spatial potential field model and a velocity potential field model to complete the initial risk screening; collision point analysis based on the initial screening results to complete the fine-tuning of risks, and generating a risk assessment value based on the risk category; local motion sampling based on vehicle state and performance constraint information to generate a set of expected candidate trajectories; calculating the performance function of each candidate trajectory based on the risk assessment value to evaluate the optimal trajectory that satisfies both risk avoidance and vehicle driving stability; and selecting the planning control command for the optimal trajectory in real time based on rolling optimization theory and sending it to the chassis controller for execution, dynamically realizing aerial risk avoidance.

[0097] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A highly agile local motion planning method for wheeled vehicles to cope with sudden aerial risks, characterized in that, The local motion planning method includes the following steps: Step 1: Initial Risk Screening; After detecting incoming aerial risks, the system quickly predicts risk information and vehicle trajectories. Based on a composite risk assessment using an artificial potential energy field, the system maps the risk location and velocity attributes to a spatial potential field model and a velocity potential field model to complete the initial risk screening. Step 2: Risk screening section; Based on the initial screening results, collision point analysis is performed to screen out the predicted aerial risk trajectories that will hit the vehicle, thus completing the risk fine screening. The risk threat coefficient is generated by combining the risk category coefficient. Step 3: Candidate trajectory sampling; Based on vehicle status and performance constraint information, local expected trajectories are sampled to generate a set of feasible candidate trajectories for vehicles within the prediction domain. Step 4: Optimal trajectory evaluation; The performance evaluation function value of each candidate trajectory is calculated based on the risk threat coefficient, and the optimal trajectory that satisfies risk avoidance and vehicle driving stability is evaluated. Step 5: Generate the optimal planning trajectory; Based on the rolling optimization theory, the optimal trajectory planning instruction is selected in real time and sent to the chassis controller for execution, thereby dynamically realizing aerial risk avoidance. Through the above steps, the highly agile wheeled vehicle local motion planning method for responding to sudden aerial risks can automatically assess aerial risks and calculate risk avoidance trajectories in real time.

2. The high-agility wheeled vehicle local motion planning method for responding to sudden aerial risks as described in claim 1, characterized in that, In the initial risk screening part of step 1, the artificial potential field adopts a coupled potential field model based on the spatial potential field model and the velocity potential field model. The spatial potential field model takes the coordinates of aerial risks and vehicle coordinates as the research objects. It analyzes whether the danger threshold is reached at a future time. If it is reached, it is determined that the aerial risk will pose a threat to the vehicle at a future time, and the aerial risk is the threat target. The velocity potential field model takes the air risk and the vehicle's current velocity vector as the research object; it determines whether the angle between the current composite velocity vector and the risk pointing vector is within the expansion angle, and if so, it is determined to be a risk target.

3. The high-agility wheeled vehicle local motion planning method for responding to sudden aerial risks as described in claim 1, characterized in that, In the risk screening section of step 2, risk category coefficients are obtained by mapping and evaluating the predicted aerial risk trajectory based on the identified target category.

4. The high-agility wheeled vehicle local motion planning method for responding to sudden aerial risks as described in claim 3, characterized in that, In the air risk and threat determination section of steps 1 and 2, the air risk prediction trajectory point sequence and the vehicle prediction trajectory point sequence are input. After Euclidean distance coarse screening, vehicle coordinate system transformation and coordinate threshold fine screening, the air threat and risk determination at the current moment is completed based on the spatial field of view model and the velocity field of view model. If it is determined that an aerial threat will hit the vehicle, output the remaining hit time, the coordinates of the hit point, and the velocity vector parameters of the hit point to subsequent steps; If it is determined that the vehicle will not be hit, the time of impact and the predicted landing point coordinates will be output to subsequent steps. If the incoming aerial object has not yet landed, the process proceeds to the next sampling time to reassess the aerial threat under the new conditions.

5. The high-agility wheeled vehicle local motion planning method for responding to sudden aerial risks as described in claim 1, characterized in that, In the aerial risk threat assessment section of step 2, if it is determined that an aerial threat will hit the vehicle, the risk threat coefficient is calculated based on the risk category coefficient corresponding to the identified target category and the remaining hit time, and then output to the subsequent steps for optimal trajectory assessment.

6. The high-agility wheeled vehicle local motion planning method for responding to sudden aerial risks as described in claim 1, characterized in that, In the local expected trajectory sampling process of step 3, trajectory points are sampled according to the vehicle's own performance. The sampling constraints include the vehicle's performance constraints, the vehicle's geometric and mechanical constraints, the vehicle's sideslip dynamics constraints, and the vehicle's obstacle avoidance safety constraints.

7. The high-agility wheeled vehicle local motion planning method for responding to sudden aerial risks as described in claim 1, characterized in that, In the optimal trajectory evaluation process of step 4, the performance evaluation function includes a target orientation evaluation term, a velocity evaluation term, and an obstacle distance evaluation term that take into account the target type.

8. The high-agility wheeled vehicle local motion planning method for responding to sudden aerial risks as described in claim 7, characterized in that, The obstacle distance evaluation item is related to the risk threat coefficient.

9. The high-agility wheeled vehicle local motion planning method for responding to sudden aerial risks as described in claim 1, characterized in that, In the process of selecting the optimal trajectory planning instruction in step 5, within each sampling interval, the top 3 control instructions of the optimal trajectory with the highest performance evaluation function value are selected and sent to the chassis controller for platform maneuver control.

10. The high-agility wheeled vehicle local motion planning method for responding to sudden aerial risks as described in claim 1, characterized in that, The method enables emergency trajectory planning in response to sudden aerial risks, improving the agility, intelligence, and safety of vehicle maneuverability, and enhancing the vehicle's survivability.

Images