A control method, device and electronic equipment for unmanned equipment

Abstract

This specification discloses a control method, device and electronic device for unmanned equipment. In the embodiments of this specification, the interaction point between the target obstacle and the unmanned equipment is first determined, and the driving strategy adopted by the target obstacle to pass through the interaction point considering the motion state of the unmanned equipment is predicted based on the motion state of the target obstacle at the current moment. The first control parameter adopted by the target obstacle to go to the interaction point according to the predicted driving strategy is determined, and the second control parameter of the unmanned equipment itself is determined based on the first control parameter, and the unmanned equipment is controlled based on the second control parameter. In this method, the target obstacle and the unmanned equipment continuously interact in the process of going to the interaction point, so as to adjust their respective control parameters. In this way, the risk of collision between the unmanned equipment and the target obstacle can be reduced, and the safety of the unmanned equipment can be improved.

Description

Technical Field

[0001] This specification relates to the field of unmanned driving, and in particular to a control method, device and electronic equipment for unmanned equipment. Background Art

[0002] In the field of unmanned driving, unmanned equipment needs to determine its own motion state according to changes in its surrounding environment in order to control the safe driving of the unmanned equipment.

[0003] In the prior art, when an unmanned device is in a scene such as an intersection or lane merging, the future motion state of other vehicles can be predicted based on the historical motion state of other vehicles around the unmanned device. Then, based on the future motion state of other vehicles, the motion state of the unmanned device itself is determined. The motion state may include: acceleration, speed, driving direction, position, etc.

[0004] However, when predicting the future motion state of other vehicles, the unmanned equipment only considers the historical motion state of other vehicles, and does not consider the interaction between the unmanned equipment and other vehicles, which leads to inaccurate prediction of the future motion state of other vehicles, causing the unmanned equipment to determine its own motion state inaccurately, and ultimately reducing the safety of the unmanned equipment's driving. Summary of the invention

[0005] The embodiments of this specification provide a control method, device and electronic device for unmanned equipment to partially solve the problems existing in the above-mentioned prior art.

[0006] The embodiments of this specification adopt the following technical solutions:

[0007] This specification provides a control method for unmanned equipment, including:

[0008] Determine a target obstacle that interacts with the unmanned device, and predict a location point where the target obstacle interacts with the unmanned device as an interaction point;

[0009] According to the motion state of the target obstacle at the current moment, predicting the driving strategy adopted by the target obstacle to pass through the interaction point in consideration of the motion state of the unmanned equipment at the current moment;

[0010] Determining a first control parameter used when the target obstacle moves toward the interaction point according to the driving strategy;

[0011] According to the first control parameter, a second control parameter for the unmanned device passing through the interaction point is determined, and based on the second control parameter, the unmanned device is controlled.

[0012] Optionally, a target obstacle that interacts with the unmanned device is determined, specifically including:

[0013] Determine a planned path corresponding to the unmanned equipment and a predicted path corresponding to each obstacle within a preset range of the unmanned equipment;

[0014] For each obstacle, determining whether there is an intersection between the planned path and the predicted path corresponding to the obstacle;

[0015] Based on the judgment results, the target obstacles that interact with the unmanned equipment are determined.

[0016] Optionally, according to the judgment result, a target obstacle having an interactive behavior with the unmanned device is determined, specifically including:

[0017] If there is an intersection between the planned path and the predicted path corresponding to the obstacle, judging whether the obstacle is an obstacle that interacts with the unmanned device according to the time it takes for the unmanned device to pass through the intersection and the time it takes for the obstacle to pass through the intersection;

[0018] If so, the obstacle is determined as a target obstacle that interacts with the unmanned device.

[0019] Optionally, according to the motion state of the target obstacle at the current moment, predicting the driving strategy adopted by the target obstacle to pass through the interaction point in consideration of the motion state of the unmanned equipment at the current moment specifically includes:

[0020] The predicted motion state of the target obstacle at the current moment predicted at the previous moment and the actual motion state of the target obstacle at the current moment are input into the prediction model, so as to determine the reward corresponding to each driving strategy according to the difference between the predicted motion state and the actual motion state through the prediction model, and predict the driving strategy adopted by the target obstacle to pass through the interaction point in consideration of the motion state of the unmanned equipment at the current moment according to the reward corresponding to each driving strategy.

[0021] Optionally, determining a reward corresponding to each driving strategy according to the difference between the predicted motion state and the actual motion state specifically includes:

[0022] For each driving strategy, according to the difference between the predicted motion state corresponding to the driving strategy and the actual motion state, at least one reward indicator corresponding to the driving strategy is determined, and the at least one reward indicator includes: a safety reward indicator, a speed reward indicator, and a position reward indicator;

[0023] A reward corresponding to the driving strategy is determined according to the at least one reward indicator.

[0024] Optionally, according to the reward corresponding to each driving strategy, predicting the driving strategy adopted by the target obstacle through the interaction point in consideration of the motion state of the unmanned device at the current moment, specifically includes:

[0025] Determining the conditional probability of the target obstacle executing the real driving strategy according to the reward corresponding to each driving strategy, wherein the real driving strategy refers to the driving strategy adopted by the target obstacle to reach the real motion state;

[0026] Determining the actual probability of the target obstacle executing each driving strategy at the current moment according to the conditional probability and the predetermined initial probability of the target obstacle executing each driving strategy;

[0027] According to the actual probability of the target obstacle executing each driving strategy at the current moment, the driving strategy adopted by the target obstacle to pass through the interaction point is predicted while considering the motion state of the unmanned equipment at the current moment.

[0028] Optionally, the method further comprises:

[0029] According to the actual probability of the target obstacle executing each driving strategy at the current moment, the initial probability of the target obstacle executing each driving strategy is updated, and the updated probability is re-determined as the initial probability, so as to predict the driving strategy adopted by the target obstacle to pass through the interaction point in the future according to the re-determined initial probability.

[0030] Optionally, determining the initial probability of the target obstacle executing each driving strategy specifically includes:

[0031] When the target obstacle is initially determined to be an obstacle that interacts with the unmanned equipment, the initial probability of the target obstacle executing each driving strategy is determined based on the time it takes for the target obstacle to pass through the interaction point, the time it takes for the unmanned equipment to pass through the interaction point, and preset traffic rules.

[0032] Optionally, determining a first control parameter used when the target obstacle moves toward the interaction point according to the driving strategy specifically includes:

[0033] If the driving strategy is a yield driving strategy, determining a traffic difference corresponding to the target obstacle according to a difference between a time length for the target obstacle to pass through the interaction point when executing the driving strategy and a time length for the target obstacle to pass through the interaction point normally;

[0034] Determine the spacing difference corresponding to the target obstacle according to the distance difference between the target obstacle and the interaction point after the target obstacle executes the driving strategy target duration; wherein the target duration is the duration for the unmanned equipment to pass through the interaction point normally;

[0035] With the goal of minimizing the traffic difference and the spacing difference, a first control parameter used in the process of the target obstacle moving toward the interaction point according to the driving strategy is determined.

[0036] Optionally, determining a first control parameter used when the target obstacle moves toward the interaction point according to the driving strategy specifically includes:

[0037] If the driving strategy is a leading driving strategy, the preset control parameter is determined as the first control parameter used in the process of the target obstacle moving toward the interaction point according to the driving strategy.

[0038] This specification provides a control device for unmanned equipment, including:

[0039] A first determination module is used to determine a target obstacle that interacts with the unmanned device, and predict a location point where the target obstacle interacts with the unmanned device as an interaction point;

[0040] A prediction module, configured to predict, based on the motion state of the target obstacle at the current moment, a driving strategy adopted by the target obstacle to pass through the interaction point in consideration of the motion state of the unmanned equipment at the current moment;

[0041] A second determination module is used to determine a first control parameter used in a process in which the target obstacle moves to the interaction point according to the driving strategy;

[0042] The control module is used to determine a second control parameter for the unmanned device passing through the interaction point according to the first control parameter, and control the unmanned device based on the second control parameter.

[0043] This specification provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the control method of the unmanned equipment described above is implemented.

[0044] The present specification provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned control method for unmanned equipment when executing the program.

[0045] At least one of the above technical solutions adopted in the embodiments of this specification can achieve the following beneficial effects:

[0046] In the embodiments of this specification, the interaction point between the target obstacle and the unmanned device is first determined, and based on the motion state of the target obstacle at the current moment, the driving strategy adopted by the target obstacle to pass through the interaction point considering the motion state of the unmanned device is predicted. The first control parameter adopted by the target obstacle to go to the interaction point according to the predicted driving strategy is determined, and based on the first control parameter, the second control parameter of the unmanned device itself is determined, and the unmanned device is controlled based on the second control parameter. In this method, the target obstacle and the unmanned device continuously interact in the process of going to the interaction point, so as to adjust their respective control parameters, so that the risk of collision between the unmanned device and the target obstacle can be reduced, and the safety of the unmanned device driving can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The drawings described herein are used to provide a further understanding of this specification and constitute a part of this specification. The illustrative embodiments and descriptions of this specification are used to explain this specification and do not constitute an improper limitation on this specification. In the drawings:

[0048] Figure 1 A schematic diagram of a control method flow chart of an unmanned device provided in an embodiment of this specification;

[0049] Figure 2 A schematic diagram of an interactive scene using an intersection as an example provided in an embodiment of this specification;

[0050] Figure 3 A schematic diagram of the structure of a control device for unmanned equipment provided in an embodiment of this specification;

[0051] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this specification. DETAILED DESCRIPTION

[0052] In the field of unmanned driving, one of the core aspects of unmanned driving technology is to accurately collect road environment information and make reasonable driving decisions accordingly. Among them, the behavior of unmanned vehicles needs to take into account the status of other vehicles (speed, direction, position, acceleration, etc.). Therefore, it is necessary to predict the movement paths of other vehicles in the future. When the movement paths of other vehicles may conflict with unmanned vehicles, unmanned vehicles need to follow traffic rules and imitate the driving habits of human drivers to design their own driving modes. Improve the stability and flexibility of unmanned vehicles while avoiding the worst risk of vehicle conflicts.

[0053] At intersections or during lane merging, the unmanned vehicle and other vehicles are in a strong interaction scenario, that is, the behaviors and intentions of the unmanned vehicle and other vehicles influence each other.

[0054] In a strong interaction scenario, in the prior art, the movement paths of other vehicles in the future are predicted based on the historical information of other vehicles, and then the unmanned vehicle responds to the predicted movement paths of other vehicles, that is, the unmanned vehicle adjusts its own driving behavior.

[0055] However, the existing technology integrates the short-term historical information of other vehicles into the instantaneous information at the current moment through calculation, which is equivalent to making static scene decisions at a higher frequency. In this way, the unmanned vehicle can only respond to the behavior of other vehicles passively, and cannot take into account the feasible strategy changes of the drivers of other vehicles, making the driving behavior of the unmanned vehicle lack flexibility and initiative. In addition, the unmanned vehicle in the existing technology cannot effectively consider the driving style of the driving behavior of other vehicles, and it is easy for conservative other vehicles to wait for each other, and it is easy for aggressive other vehicles to have a collision risk.

[0056] The control method of the unmanned equipment provided in this specification aims to predict the driving strategies of other vehicles interacting with the unmanned equipment through a Bayesian online learning model, and then the unmanned equipment determines its own driving strategy based on the predicted driving strategies of other vehicles.

[0057] In order to make the purpose, technical solutions and advantages of this specification more clear, the technical solutions of this specification will be clearly and completely described below in combination with the specific embodiments of this specification and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of this specification, not all of them. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this specification.

[0058] The technical solutions provided by the embodiments of this specification are described in detail below in conjunction with the accompanying drawings.

[0059] Figure 1 The schematic flow chart of the control method of the unmanned equipment provided in the embodiment of this specification includes:

[0060] S100: Determine a target obstacle that interacts with the unmanned device, and predict a location point where the target obstacle interacts with the unmanned device as an interaction point.

[0061] In the embodiments of this specification, the unmanned equipment may be an unmanned vehicle, that is, an unmanned car, and the unmanned equipment may be applied to the field of logistics and distribution, including both instant delivery fields such as takeout and delivery, and other non-instant delivery fields. It may also be applied to manned services.

[0062] In the embodiments of this specification, the unmanned device can obtain obstacles and the motion state of each obstacle within the preset range of the unmanned device through sensors installed on the unmanned device. The sensors may include: visual sensors, laser radars, sensors for positioning, etc. The motion state of the obstacle may include: speed, driving direction, longitude and latitude position, lane position, acceleration, etc. The obstacle may be a manned vehicle.

[0063] After obtaining each obstacle within the preset range of the unmanned device, the predicted paths corresponding to each obstacle are predicted based on the historical motion states of each obstacle obtained by the unmanned device. Then, based on the predicted paths corresponding to each obstacle, the planned paths corresponding to the unmanned device are planned for the unmanned device. Then, based on the planned paths corresponding to the unmanned device and the predicted paths corresponding to each obstacle, obstacles that interact with the unmanned device are selected from the obstacles within the preset range of the unmanned device as target obstacles. And the location point where the target obstacle interacts with the unmanned device is predicted as the interaction point. Among them, the target obstacle that interacts with the unmanned device refers to an obstacle that conflicts with the unmanned device during driving. The location point where there is a conflict between the unmanned device and the target obstacle, that is, the location point where the target obstacle interacts with the unmanned device, is the interaction point.

[0064] When selecting a target obstacle that interacts with the unmanned device from obstacles within the preset range of the unmanned device, the planned path corresponding to the unmanned device and the predicted path corresponding to each obstacle within the preset range of the unmanned device can be obtained. For the predicted path corresponding to each obstacle, it is determined whether there is an intersection between the planned path and the predicted path corresponding to the obstacle. According to the judgment result, the target obstacle that interacts with the unmanned device is determined from the obstacles.

[0065] Specifically, if there is no intersection between the planned path and the predicted path corresponding to the obstacle, it is determined that the obstacle is not an obstacle that interacts with the unmanned device.

[0066] If there is an intersection between the planned path and the predicted path corresponding to the obstacle, it is determined whether the obstacle is a target obstacle that interacts with the unmanned device based on the time it takes the unmanned device to pass through the intersection and the time it takes the obstacle to pass through the intersection.

[0067] When judging whether the obstacle is a target obstacle with which the unmanned device interacts based on the time it takes for the unmanned device to pass through the intersection and the time it takes for the obstacle to pass through the intersection, the obstacle can be judged whether it is an obstacle with which the unmanned device interacts based on the difference between the time it takes for the unmanned device to pass through the intersection and the time it takes for the obstacle to pass through the intersection, and a judgment result of the obstacle is obtained. If the difference between the time it takes for the unmanned device to pass through the intersection and the time it takes for the obstacle to pass through the intersection is not greater than a preset first time difference threshold, it is determined that the obstacle is a target obstacle with which the unmanned device interacts. Based on the judgment result of each obstacle, an obstacle with which the unmanned device interacts is selected from each obstacle as a target obstacle.

[0068] In addition, when judging whether the obstacle is a target obstacle with which the unmanned device interacts based on the time it takes for the unmanned device to pass through the intersection and the time it takes for the obstacle to pass through the intersection, the first difference between the time it takes for the unmanned device to pass through the intersection and the time it takes for the obstacle to reach the intersection can also be determined. At the same time, the second difference between the time it takes for the obstacle to pass through the intersection and the time it takes for the unmanned device to reach the intersection can be determined. Then, from the first difference and the second difference, the smallest difference is selected as the interaction time difference. Finally, based on the interaction time difference, it is judged whether the obstacle is an obstacle with which the unmanned device interacts, and a judgment result corresponding to the obstacle is obtained. If the interaction time difference is not greater than the preset second time difference threshold, it is determined that the obstacle is a target obstacle with which the unmanned device interacts. Based on the judgment result of each obstacle, an obstacle with which the unmanned device interacts is selected from each obstacle as a target obstacle.

[0069] Among them, when determining the time it takes for the unmanned equipment to pass through the intersection, the time it takes for the unmanned equipment to pass through the intersection can be determined based on the length of the unmanned equipment, the distance between the front of the unmanned equipment and the intersection along the planned path, the width of the obstacle, and the instantaneous speed of the unmanned equipment.

[0070] The specific formula is: Among them, T 11 It represents the time it takes for the unmanned equipment to pass the intersection, s1 represents the distance from the front of the unmanned equipment to the intersection along the planned path, l1 represents the length of the unmanned equipment, w2 represents the width of the obstacle, and v1 represents the instantaneous speed of the unmanned equipment.

[0071] When determining the time it takes for the unmanned equipment to reach the intersection, the time it takes for the unmanned equipment to reach the intersection can be determined based on the distance between the front of the unmanned equipment and the intersection along the planned path and the instantaneous speed of the unmanned equipment.

[0072] The specific formula is: Among them, T 12 Indicates the time it takes for the unmanned equipment to reach the intersection.

[0073] When determining the time it takes for the obstacle to pass through the intersection, the time it takes for the obstacle to pass through the intersection can be determined based on the length of the obstacle's body, the distance between the front of the obstacle and the intersection along the predicted path, the body width of the unmanned equipment, and the instantaneous speed of the obstacle.

[0074] The specific formula is: Among them, T 21 It represents the time it takes for the obstacle to pass through the intersection, s2 represents the distance from the front of the obstacle to the intersection along the predicted path, l2 represents the length of the obstacle, w1 represents the width of the unmanned equipment, and v2 represents the instantaneous speed of the obstacle.

[0075] When determining the time it takes for the obstacle to reach the intersection, the time it takes for the obstacle to reach the intersection can be determined based on the distance between the front of the obstacle and the intersection along the predicted path of the obstacle and the instantaneous speed of the obstacle.

[0076] The specific formula is: Among them, T 22 Indicates the time it takes for the obstacle to reach the intersection point.

[0077] The formula for the first difference is: T 11 -T 22 , the formula for the second difference is: T 21 -T 12 .

[0078] After determining the target obstacle that has an interactive behavior with the unmanned device, when predicting the interaction point where the target obstacle interacts with the unmanned device, the location point where the target obstacle interacts with the unmanned device can be predicted based on the planned path of the unmanned device and the predicted path of the target obstacle as the interaction point, that is, the intersection point between the planned path of the unmanned device and the predicted path of the target obstacle.

[0079] Based on the above description of determining the target obstacle and the interaction point, the embodiment of this specification provides a schematic diagram of an interaction scene taking an intersection as an example, such as Figure 2 shown.

[0080] exist Figure 2 In the figure, the planned path of the unmanned equipment is L1, the length of the unmanned equipment is l1, the width of the equipment is w1, and the distance from the front of the vehicle to the interaction point is s1. The predicted path of the target obstacle is L2, the length of the target obstacle is l2, the width of the equipment is w2, and the distance from the front of the vehicle to the interaction point is s2.

[0081] In addition, when determining the interaction point between the unmanned equipment and the target obstacle, the position points where the unmanned equipment is traveling along the planned path and the target obstacle is traveling along the predicted path can be determined as interaction points based on the body length of the unmanned equipment, the body width of the obstacle and the instantaneous speed of the unmanned equipment, as well as the body length of the target obstacle, the body width of the unmanned equipment and the instantaneous speed of the target obstacle.

[0082] It should be noted that Figure 1 The control method of the unmanned equipment shown can be applied to strong interactive scenes such as intersections and merging lanes, and can also be applied to simple road scenes. Figure 1 The control method of the unmanned equipment shown can be applied to the unmanned equipment, and can also be applied to the server that controls the unmanned equipment.

[0083] S102: According to the current motion state of the target obstacle, predict the driving strategy adopted by the target obstacle to pass through the interaction point in consideration of the current motion state of the unmanned equipment.

[0084] In the embodiment of the present specification, after the target obstacle is determined, the driving strategy adopted by the target obstacle through the interaction point in consideration of the current motion state of the unmanned device can be predicted based on the current motion state of the target obstacle. That is, when the target obstacle detects that there is a conflict between the unmanned device and its own driving path, and the target obstacle needs to determine its own driving strategy based on the current motion state of the unmanned device, the unmanned device predicts the driving strategy adopted by the target obstacle at the current moment.

[0085] In this way, the game behavior between the target obstacle and the unmanned device can be determined according to the motion state of the target obstacle at the current moment and the motion state of the unmanned device at the current moment, that is, the interactive behavior between the target obstacle and the unmanned device can be determined.

[0086] The motion state may include: speed, acceleration, position, driving direction, etc. The driving strategy may include: first-come-first-served driving strategy and yielding driving strategy.

[0087] In the case where the target obstacle is not initially determined as an obstacle that has an interactive behavior with the unmanned device, the driving strategy adopted by the target obstacle at the current moment by taking into account the motion state of the unmanned device at the current moment can be predicted based on the predicted motion state of the target obstacle at the previous moment predicted by the unmanned device and the actual motion state of the target obstacle at the current moment. The predicted motion state may include: predicted speed, predicted acceleration, predicted position, predicted driving direction, etc., and the actual motion state may include: actual speed, actual acceleration, actual position, actual driving direction, etc.

[0088] Specifically, the predicted motion state of the target obstacle at the current moment predicted by the unmanned equipment at the previous moment and the actual motion state of the target obstacle at the current moment can be input into the prediction model, so as to determine the reward corresponding to each driving strategy according to the difference between the predicted motion state and the actual motion state through the prediction model. Determining the reward corresponding to each driving strategy refers to determining the preference coefficient of the target obstacle for each driving strategy.

[0089] The unmanned device can predict the driving strategy adopted by the target obstacle through the interaction point based on the reward corresponding to each driving strategy of the target obstacle (i.e., based on the preference coefficient of the target obstacle for each driving strategy), taking into account the current motion state of the unmanned device; wherein the prediction model includes: a Bayesian online learning model or a Bayesian online learning model combined with reinforcement learning. The prediction model can learn the driving style preferred by the target obstacle based on the predicted motion state of the target obstacle input in real time and the actual motion state of the target obstacle, and the driving style may include: executing a yield driving strategy and executing a first-come-first-served driving strategy.

[0090] When determining the reward corresponding to each driving strategy according to the difference between the predicted motion state and the actual motion state, for each driving strategy, at least one reward index (i.e., preference index) corresponding to the driving strategy can be determined according to the difference between the predicted motion state and the actual motion state corresponding to the driving strategy, wherein the at least one reward index includes: a safety reward index (i.e., a safety preference index), a speed reward index (i.e., a speed preference index), and a position reward index (i.e., a position preference index). Then, according to the at least one reward index corresponding to the driving strategy, the reward corresponding to the driving strategy is determined, that is, the preference coefficient of the target obstacle for the driving strategy.

[0091] Among them, the safety reward index refers to the minimum acceleration between the predicted acceleration in the predicted motion state of the target obstacle and the acceleration of the motion state of the unmanned equipment at the current moment. The speed reward index refers to the difference between the predicted speed and the actual speed. The position reward index refers to the difference between the predicted position and the actual position.

[0092] The formula for the safety reward indicator is: Among them, v p Expressed as the predicted speed, s p Represented as predicted position. Represented as the predicted acceleration of the target obstacle. It is expressed as the acceleration of the unmanned equipment in the current state of motion, v o Indicates the speed of the unmanned equipment in the current state of motion, s o Indicates the position of the unmanned equipment in its current state of motion.

[0093] The formula for the speed reward indicator is: Re v =v r -v p Among them, v r Indicates the true speed.

[0094] The formula for the position reward indicator is: Re l =s r -s p Among them, s r Indicates the real position.

[0095] When determining the reward corresponding to the driving strategy based on at least one reward indicator corresponding to the driving strategy, the weight corresponding to each reward indicator corresponding to the driving strategy can be determined, and then, based on the weight corresponding to each reward indicator, each reward indicator can be weighted summed to obtain the reward corresponding to the driving strategy.

[0096] When determining the reward corresponding to the driving strategy according to at least one reward indicator corresponding to the driving strategy, multiple groups of reward coefficients may be obtained or sampled, and for each driving strategy corresponding to each group of reward coefficients, the reward corresponding to the group of reward coefficients for the driving strategy may be determined according to the group of reward coefficients and at least one reward indicator corresponding to the driving strategy. For each driving strategy, each group of reward coefficients may represent the preference weight of the target obstacle for the driving strategy, that is, the weight of each reward indicator corresponding to the driving strategy, that is, the preference weight of the target obstacle for each preference indicator corresponding to the driving strategy.

[0097] Specifically, based on the sub-coefficients in the group of reward coefficients, at least one reward indicator corresponding to the driving strategy is weighted and summed to obtain the reward corresponding to the driving strategy by the group of reward coefficients. When the sub-coefficients in a group of reward coefficients obey a normal distribution, each group of reward coefficients also obeys a normal distribution, and further, the reward corresponding to the driving strategy by each group of reward coefficients also obeys a normal distribution.

[0098] Among them, a group of reward coefficients may include sub-coefficients for reward indicators, and each group of reward coefficients includes different sub-coefficients. The sub-coefficients for reward indicators can represent the weights of reward indicators, and the sub-coefficients for reward indicators obey multivariate normal distribution, and the number of elements of the normal distribution is determined by the number of reward indicators. In addition, different driving strategies correspond to different reward indicators.

[0099] For example, the multiple sets of reward coefficients obtained can be θ1,…,θ n is n groups of reward coefficients, θ n is the nth group of reward coefficients. A group of reward coefficients contains sub-coefficients for reward indicators, with θ n For example, when the number of reward indicators is 2, θ n It follows a bivariate normal distribution. When the number of reward indicators is 3, θ n It obeys the ternary normal distribution, and so on. When the number of reward indicators is 3, the sub-coefficients in the nth group of reward coefficients can be: θ n =(θ n 1 ,θ n 2 ,θ n 3 ).θ n 1 Represented as the first sub-coefficient in this set of reward coefficients.

[0100] For each driving strategy, assume that the reward index corresponding to this driving strategy is the safety reward index Re s , speed reward index Re v , Position reward index Re l , then θ n The reward corresponding to this driving strategy is: Re(I p |θ n )=θ n 1 *Re s +θ n 2 *Re v +θ n 3 *Re l Among them, I p Represents a driving strategy, and in this example represents this driving strategy.

[0101] After determining the reward corresponding to each driving strategy, the driving strategy adopted by the target obstacle to pass through the interaction point in consideration of the current motion state of the unmanned device can be predicted based on the reward corresponding to each driving strategy. That is, based on the preference coefficient of the target obstacle to each driving strategy, the driving strategy adopted by the target obstacle to pass through the interaction point in consideration of the current motion state of the unmanned device is predicted.

[0102] Specifically, the conditional probability of the target obstacle executing the real driving strategy can be determined based on the reward corresponding to each driving strategy, where the real driving strategy refers to the driving strategy adopted by the target obstacle to reach the real motion state, which here refers to the driving strategy adopted by the target obstacle at the previous moment. Then, based on the conditional probability and the predetermined initial probability of the target obstacle executing each driving strategy, the posterior probability of the target obstacle executing each driving strategy at the current moment is determined as the actual probability of the target obstacle executing each driving strategy at the current moment. Based on the actual probability of the target obstacle executing each driving strategy at the current moment, the driving strategy adopted by the target obstacle to pass the interaction point is predicted taking into account the motion state of the unmanned equipment at the current moment.

[0103] The initial probability of the target obstacle executing each driving strategy may be the probability of the target obstacle executing each driving strategy at the last moment determined at the last moment, or the probability of the target obstacle executing each driving strategy at the beginning. In addition, the probability of the target obstacle executing each driving strategy at the last moment determined at the last moment is, that is, the probability of the target obstacle executing each driving strategy at the last moment after being updated at the last moment.

[0104] The formula for calculating the conditional probability of the target obstacle executing the real driving strategy is: Re(I r |θ n ) represents the reward for executing the real driving strategy for the target obstacle, ∑ I Re(I|θ n ) represents the reward for executing all driving strategies for the target obstacle.

[0105] In order to improve the accuracy of the driving strategy predicted for the target obstacle at the current moment, the driving strategies that the target obstacle has adopted in the past can be combined to predict the driving strategy that the target obstacle may adopt at the current moment.

[0106] When determining the actual probability of the target obstacle executing each driving strategy at the current moment according to the conditional probability and the predetermined initial probability of the target obstacle executing each driving strategy, the actual probability of the target obstacle executing each driving strategy at the current moment can be determined according to the conditional probability corresponding to the target obstacle executing the real driving strategy at all historical moments (including the last moment) and the probability of the target obstacle executing each driving strategy at the last moment determined at the last moment. The "real driving strategy" here refers to the driving strategy adopted by the target obstacle at all historical moments.

[0107] The formula for calculating the actual probability of the target obstacle executing each driving strategy at the current moment is: Among them, P(θ n |I r ) is the target obstacle at the current time t0 in θ n The posterior probability of executing each driving strategy under the reward coefficient is expressed as all historical moments from t0-R+1 to t0, The conditional probability of executing the real driving strategy at all historical moments is used to approximate that the target obstacle is at θ at the current moment. n The prior probability of executing a realistic driving strategy under the reward coefficient. P(θ n ) represents the target obstacle determined at the last moment at θ n The probability of executing each driving strategy under the reward coefficient, that is, the initial probability. In addition, when θ is different, P(θ n |I r ), P(θ n ) are also different.

[0108] When predicting the driving strategy adopted by the target obstacle when passing through the interaction point taking into account the motion state of the unmanned equipment at the current moment based on the actual probability of the target obstacle executing each driving strategy at the current moment, the driving strategy with the highest actual probability can be determined as the driving strategy that the target obstacle may adopt when passing through the interaction point taking into account the motion state of the unmanned equipment at the current moment.

[0109] In addition, in the case of obtaining multiple sets of reward coefficients, after determining the reward corresponding to each set of reward coefficients for each driving strategy, the conditional probability of the target obstacle corresponding to each set of reward coefficients executing the actual driving strategy can be determined based on the reward corresponding to each set of reward coefficients for each driving strategy. For each set of reward coefficients, the actual probability of the target obstacle corresponding to the set of reward coefficients executing each driving strategy at the current moment is determined based on the conditional probability corresponding to the set of reward coefficients and the initial probability of the target obstacle corresponding to the set of reward coefficients executing each driving strategy. Based on the actual probability of the target obstacle corresponding to each set of reward coefficients executing each driving strategy at the current moment, the driving strategy adopted by the target obstacle through the interaction point is predicted taking into account the motion state of the unmanned device at the current moment.

[0110] Since each set of reward coefficients corresponds to the actual probability of the target obstacle executing each driving strategy at the current moment, and each set of reward coefficients obeys a normal distribution, therefore, for each driving strategy executed by the target obstacle at the current moment, the actual probability of each driving strategy actually obeys a normal distribution.

[0111] Therefore, when predicting the driving strategy that the target obstacle may adopt at the current moment, for each driving strategy, the probability of the reward corresponding to the driving strategy being at the average reward can be determined as the actual probability corresponding to the target obstacle adopting the driving strategy at the current moment based on the actual probability distribution of the driving strategy and the reward of the driving strategy. Finally, based on the actual probability corresponding to the target obstacle adopting each driving strategy at the current moment, the driving strategy that the target obstacle may adopt at the current moment is predicted, that is, the driving strategy adopted by the target obstacle when passing through the interaction point is predicted, taking into account the motion state of the unmanned device at the current moment. In other words, the driving strategy with the highest actual probability is determined as the driving strategy adopted by the target obstacle when passing through the interaction point, taking into account the motion state of the unmanned device at the current moment.

[0112] In addition, the initial probability of the target obstacle executing each driving strategy can be updated according to the actual probability of the target obstacle executing each driving strategy at the current moment, and the updated probability can be redetermined as the initial probability, so as to predict the driving strategy adopted by the subsequent target obstacle through the interaction point according to the redetermined initial probability.

[0113] When multiple sets of reward coefficients are obtained, the initial probability of the target obstacle executing each driving strategy obeys the normal distribution. Therefore, the initial probability distribution of the target obstacle executing each driving strategy at the current moment can be updated, and the updated probability distribution can be re-determined as the initial probability distribution, so as to predict the driving strategy adopted by the subsequent target obstacle passing through the interaction point according to the re-determined initial probability distribution.

[0114] S104: Determine a first control parameter used when the target obstacle moves toward the interaction point according to the driving strategy.

[0115] S106: Determine a second control parameter for the unmanned device passing through the interaction point according to the first control parameter, and control the unmanned device based on the second control parameter.

[0116] In an embodiment of the present specification, after predicting the driving strategy that the target obstacle adopts with the highest probability at the current moment, the driving strategy of the unmanned device itself can be determined according to the predicted driving strategy of the target obstacle. Specifically, the first control parameter adopted by the target obstacle in the process of going to the interaction point according to the predicted driving strategy can be predicted (determined) first. Then, the unmanned device determines the second control parameter for the unmanned device itself to pass through the interaction point based on the first control parameter, and controls the unmanned device based on the second control parameter to make the unmanned device move. Among them, the first control parameter and the second control parameter can be acceleration, speed, position, direction, etc. If the driving strategy of the target obstacle is a yield driving strategy, the driving strategy of the unmanned device is a priority driving strategy. Conversely, if the driving strategy of the target obstacle is a priority driving strategy, the driving strategy of the unmanned device is a yield driving strategy.

[0117] The first case: When the driving strategy of the target obstacle is the advance driving strategy, the unmanned device predicts that the target obstacle will travel at a constant speed or accelerate during the execution of the advance driving strategy. In this case, the unmanned device executes the yield driving strategy, and the unmanned device decelerates or stops during the execution of the yield driving strategy.

[0118] The second case: When the driving strategy of the target obstacle is the yield driving strategy, the unmanned device predicts that the target obstacle will decelerate during the execution of the advance driving strategy. In this case, the unmanned device executes the advance driving strategy, and the unmanned device drives at a constant speed or accelerates during the execution of the advance driving strategy.

[0119] The third case: when the actual probabilities of each driving strategy adopted by the target obstacle at the current moment are relatively close, the unmanned device cannot predict which driving strategy the target obstacle adopts at the current moment. In this case, the unmanned device predicts that the target obstacle will execute a mixed driving strategy.

[0120] For the first case: if the driving strategy of the target obstacle is the advance driving strategy, the preset control parameter is determined as the first control parameter used in the process of the target obstacle moving toward the interaction point according to the predicted driving strategy.

[0121] For the second case:

[0122] If the predicted driving strategy of the target obstacle is a yield driving strategy, in order to enable the target obstacle to still efficiently and safely pass through the interaction point when the yield driving strategy is executed, it is necessary to predict the acceleration of the target obstacle during the execution of the yield driving strategy.

[0123] Specifically, the traffic difference corresponding to the target obstacle can be determined according to the difference between the time it takes for the target obstacle to pass through the interaction point by executing the predicted driving strategy and the time it takes for the target obstacle to pass through the interaction point normally. Wherein, passing through the interaction point normally may mean that the target obstacle passes through the interaction point without being affected by any vehicle, that is, the target obstacle passes through the interaction point normally when the target obstacle executes the advance driving strategy and passes through the interaction point.

[0124] The formula for calculating the traffic difference corresponding to the target obstacle is: Among them, O is the first driving strategy, Y is the first driving strategy, EL Y It is expressed as the passage difference corresponding to the target obstacle, and a is the acceleration (unknown).

[0125] Then, the spacing difference corresponding to the target obstacle is determined based on the distance difference between the target obstacle and the interaction point after the target obstacle executes the predicted driving strategy target duration. The target duration is the duration for the unmanned equipment to pass through the interaction point normally.

[0126] The formula for calculating the distance difference corresponding to the target obstacle is: Among them, CL Y It is expressed as the distance difference corresponding to the target obstacle, Indicates the time it takes for the unmanned device to pass the interaction point when executing the advance driving strategy. Indicates the distance difference between the target obstacle and the interaction point after the unmanned device passes through the interaction point.

[0127] Finally, with the goal of minimizing the traffic difference and the spacing difference, the first control parameter used by the target obstacle in the process of moving to the interaction point according to the predicted driving strategy is determined.

[0128] The total difference of traffic difference and spacing difference is: TL Y =EL Y +CL Y It can be seen that TL Y It is a functional relationship about acceleration a. With the goal of minimizing the traffic difference and the spacing difference, the acceleration a is solved to obtain the first control parameter.

[0129] For the third case:

[0130] If the unmanned equipment cannot predict which driving strategy the target obstacle adopts at the current moment, according to the actual probability of the target obstacle executing each driving strategy at the current moment, the first control parameters determined by the target obstacle executing each driving strategy at the current moment can be weighted summed to obtain the mixed control parameters of the target obstacle executing the driving strategy at the current moment.

[0131] Among them, the formula of the hybrid control parameter of the target obstacle executing the driving strategy at the current moment is: Among them, P2(O) represents the actual probability of executing the first-pass driving strategy for the target obstacle, and P2(Y) represents the actual probability of executing the yield driving strategy for the target obstacle. represents the first control parameter determined when executing the advance driving strategy for the target obstacle, which may be acceleration, It represents the first control parameter determined when executing the yield driving strategy for the target obstacle.

[0132] Through the above Figure 1 As can be seen from the method shown, this manual first determines the interaction point between the target obstacle and the unmanned device, and predicts the driving strategy adopted by the target obstacle to pass through the interaction point based on the motion state of the target obstacle at the current moment, taking into account the motion state of the unmanned device. Determine the first control parameter adopted by the target obstacle to go to the interaction point according to the predicted driving strategy, and determine the second control parameter of the unmanned device itself based on the first control parameter, and control the unmanned device based on the second control parameter. In this method, the target obstacle and the unmanned device continuously interact in the process of going to the interaction point, so as to adjust their respective control parameters. In this way, the risk of collision between the unmanned device and the target obstacle can be reduced, and the safety of the unmanned device can be improved.

[0133] Furthermore, in Figure 1In the steps S102 to S106 shown, when the target obstacle is initially determined to be an obstacle that interacts with the unmanned device, the target obstacle does not have the predicted motion state of the target obstacle at the current moment predicted at the previous moment, so the probability of the target obstacle adopting each driving strategy at the current moment can be initialized first. Then, based on the initialized probability of each driving strategy, the first control parameter of the target obstacle during the execution of each driving strategy by the target obstacle at the current moment is determined. The unmanned device determines the second control parameter of the unmanned device itself based on the first control parameter.

[0134] Specifically, when the target obstacle is initially determined to be an obstacle that interacts with the unmanned device, the duration of the target obstacle passing through the interaction point can be determined based on the target obstacle's current state of motion. Then, based on the duration of the target obstacle passing through the interaction point, the duration of the unmanned device passing through the interaction point, and the preset traffic rules, the driving strategy adopted by the target obstacle to pass through the interaction point is predicted, taking into account the current state of motion of the unmanned device. Among them, the preset traffic rules refer to the priority of the driving direction of the target obstacle and the unmanned device. The priority of the driving direction in the traffic rules is from high to low: straight > left turn > right turn, and the priority values ​​correspond to 3>2>1 respectively.

[0135] That is, according to the time it takes for the target obstacle to pass through the interaction point, the time it takes for the unmanned equipment to pass through the interaction point, and the priority of the driving direction between the target obstacle and the unmanned equipment, the initial probability of the target obstacle executing each driving strategy at the current moment is determined. Then, the driving strategy with the largest initial probability is determined as the driving strategy executed by the target obstacle at the current moment.

[0136] Among them, the calculation formula for the initial probability of the target obstacle executing the advance driving strategy is: Among them, P2(O) represents the probability of executing the first-pass driving strategy for the target obstacle, O represents the first-pass driving strategy, a represents the time weight, r represents the priority of different driving directions, r2 represents the priority of the driving direction of the target obstacle, r1 represents the priority of the driving direction of the unmanned equipment, r2-r1 represents the driving priority, and b represents the weight of the driving priority. The initial probability of the target obstacle executing the yield driving strategy is P2(Y)=1-P2(O). Y represents the yield driving strategy.

[0137] It should be noted that, among the formulas contained in this specification, the subscript of the formula used to describe the relevant content of unmanned equipment is "1", the subscript of the formula used to describe the relevant content of obstacles or target obstacles is "2", the superscript of the formula used to describe the priority driving strategy is "O", and the superscript of the formula used to describe the yield driving strategy is "Y".

[0138] The above is a control method for unmanned equipment provided in the embodiments of this specification. Based on the same idea, this specification also provides corresponding devices, storage media and electronic devices.

[0139] Figure 3 A schematic diagram of a control device for unmanned equipment provided in an embodiment of this specification, the device comprising:

[0140] The first determination module 301 is used to determine a target obstacle that interacts with the unmanned device, and predict a location point where the target obstacle interacts with the unmanned device as an interaction point;

[0141] A prediction module 302, configured to predict, based on the motion state of the target obstacle at the current moment, a driving strategy adopted by the target obstacle to pass through the interaction point in consideration of the motion state of the unmanned equipment at the current moment;

[0142] A second determination module 303 is used to determine a first control parameter used when the target obstacle moves to the interaction point according to the driving strategy;

[0143] The control module 304 is used to determine a second control parameter for the unmanned device passing through the interaction point according to the first control parameter, and control the unmanned device based on the second control parameter.

[0144] Optionally, the first determination module 301 is specifically used to determine the planned path corresponding to the unmanned equipment and the predicted path corresponding to each obstacle within a preset range of the unmanned equipment; for each obstacle, determine whether there is an intersection between the planned path and the predicted path corresponding to the obstacle; and based on the judgment result, determine the target obstacle that has interactive behavior with the unmanned equipment.

[0145] Optionally, the first determination module 301 is specifically used to determine whether the obstacle is an obstacle that interacts with the unmanned device based on the time it takes for the unmanned device to pass through the intersection and the time it takes for the obstacle to pass through the intersection if there is an intersection between the planned path and the predicted path corresponding to the obstacle; if so, determine the obstacle as a target obstacle that interacts with the unmanned device.

[0146] Optionally, the prediction module 302 is specifically used to input the predicted motion state of the target obstacle at the current moment predicted at the previous moment and the actual motion state of the target obstacle at the current moment into the prediction model, so as to determine the reward corresponding to each driving strategy according to the difference between the predicted motion state and the actual motion state through the prediction model, and predict the driving strategy adopted by the target obstacle to pass through the interaction point taking into account the motion state of the unmanned equipment at the current moment based on the reward corresponding to each driving strategy.

[0147] Optionally, the prediction module 302 is specifically used to determine, for each driving strategy, at least one reward indicator corresponding to the driving strategy based on the difference between the predicted motion state corresponding to the driving strategy and the actual motion state, the at least one reward indicator including: a safety reward indicator, a speed reward indicator, and a position reward indicator; and determine the reward corresponding to the driving strategy based on the at least one reward indicator.

[0148] Optionally, the prediction module 302 is specifically used to determine the conditional probability of the target obstacle executing a real driving strategy based on the reward corresponding to each driving strategy, wherein the real driving strategy refers to the driving strategy adopted by the target obstacle to reach the real motion state; determine the actual probability of the target obstacle executing each driving strategy at the current moment based on the conditional probability and a predetermined initial probability of the target obstacle executing each driving strategy; and predict the driving strategy adopted by the target obstacle to pass through the interaction point while considering the motion state of the unmanned equipment at the current moment based on the actual probability of the target obstacle executing each driving strategy at the current moment.

[0149] Optionally, the prediction module 302 is also used to update the initial probability of the target obstacle executing each driving strategy according to the actual probability of the target obstacle executing each driving strategy at the current moment, and redetermine the updated probability as the initial probability, so as to predict the driving strategy subsequently adopted by the target obstacle when passing through the interaction point according to the redetermined initial probability.

[0150] Optionally, the prediction module 302 is specifically used to determine the initial probability of the target obstacle executing each driving strategy according to the time it takes for the target obstacle to pass through the interaction point, the time it takes for the unmanned equipment to pass through the interaction point, and preset traffic rules when the target obstacle is initially determined to be an obstacle that interacts with the unmanned equipment.

[0151] Optionally, the second determination module 303 is specifically used to, if the driving strategy is a yield driving strategy, determine the traffic difference corresponding to the target obstacle according to the difference between the time it takes for the target obstacle to execute the driving strategy to pass the interaction point and the time it takes for the target obstacle to pass the interaction point normally; determine the spacing difference corresponding to the target obstacle according to the distance difference between the target obstacle and the interaction point after the target obstacle executes the driving strategy for the target time, wherein the target time is the time it takes for the unmanned equipment to pass the interaction point normally; and determine the first control parameter adopted by the target obstacle in the process of moving to the interaction point according to the driving strategy with the goal of minimizing the traffic difference and the spacing difference.

[0152] Optionally, the second determination module 303 is specifically configured to, if the driving strategy is a prior driving strategy, determine a preset control parameter as a first control parameter used when the target obstacle moves toward the interaction point according to the driving strategy.

[0153] This specification also provides a computer-readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, it can be used to perform the above Figure 1 A control method for unmanned equipment is provided.

[0154] based on Figure 1 The control method of the unmanned equipment shown in the present specification also provides Figure 4 The structural diagram of the electronic device shown in FIG. Figure 4 At the hardware level, the electronic device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory, and may also include other hardware required for the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to achieve the above Figure 1 The control method of the unmanned equipment.

[0155] Of course, in addition to software implementation, this specification does not exclude other implementation methods, such as logic devices or a combination of software and hardware, etc., that is to say, the executor of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0156] In the 1990s, improvements to a technology could be clearly distinguished as hardware improvements (for example, improvements to the circuit structure of diodes, transistors, switches, etc.) or software improvements (improvements to the method flow). However, with the development of technology, many improvements to the method flow today can be regarded as direct improvements to the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that an improvement in a method flow cannot be implemented using a hardware entity module. For example, a programmable logic device (PLD) (such as a field programmable gate array (FPGA)) is such an integrated circuit whose logical function is determined by the user's programming of the device. Designers can "integrate" a digital system on a PLD by programming it themselves, without having to ask a chip manufacturer to design and produce a dedicated integrated circuit chip. Moreover, nowadays, instead of manually making integrated circuit chips, this kind of programming is mostly implemented by "logic compiler" software, which is similar to the software compiler used when developing and writing programs, and the original code before compilation must also be written in a specific programming language, which is called hardware description language (HDL). There is not only one HDL, but many kinds, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. The most commonly used ones are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also know that it is only necessary to program the method flow slightly in the above-mentioned hardware description languages ​​and program it into the integrated circuit, and then it is easy to obtain the hardware circuit that implements the logic method flow.

[0157] The controller can be implemented in any appropriate manner, for example, the controller can take the form of a microprocessor or processor and a computer-readable medium storing a computer-readable program code (such as software or firmware) that can be executed by the (micro)processor, a logic gate, a switch, an application-specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320. The memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art also know that in addition to implementing the controller in a purely computer-readable program code manner, the controller can be implemented in the form of a logic gate, a switch, an application-specific integrated circuit, a programmable logic controller, and an embedded microcontroller by logically programming the method steps. Therefore, this controller can be considered as a hardware component, and the devices included therein for implementing various functions can also be regarded as structures within the hardware component. Or even, the devices for implementing various functions can be regarded as both software modules for implementing the method and structures within the hardware component.

[0158] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0159] For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0160] Those skilled in the art will appreciate that the embodiments of this specification may be provided as methods, systems, or computer program products. Therefore, this specification may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0161] This specification is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of this specification. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0162] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0163] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0164] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0165] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0166] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0167] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0168] Those skilled in the art will appreciate that the embodiments of this specification may be provided as methods, systems or computer program products. Therefore, this specification may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0169] This specification may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0170] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0171] The above description is only an embodiment of the present specification and is not intended to limit the present specification. For those skilled in the art, the present specification may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present specification shall be included in the scope of the claims of the present specification.

Claims

1. A control method for unmanned equipment, characterized in that: include: Determine a target obstacle that interacts with the unmanned device, and predict a location point where the target obstacle interacts with the unmanned device as an interaction point; According to the motion state of the target obstacle at the current moment, predicting the driving strategy adopted by the target obstacle to pass through the interaction point in consideration of the motion state of the unmanned equipment at the current moment; Determining a first control parameter used when the target obstacle moves toward the interaction point according to the driving strategy; Determining, according to the first control parameter, a second control parameter for the unmanned device passing through the interaction point, and controlling the unmanned device based on the second control parameter; The first control parameter used in the process of determining that the target obstacle moves to the interaction point according to the driving strategy specifically includes: If the driving strategy is a yield driving strategy, determining a traffic difference corresponding to the target obstacle according to a difference between a time length for the target obstacle to pass through the interaction point when executing the driving strategy and a time length for the target obstacle to pass through the interaction point normally; Determine the spacing difference corresponding to the target obstacle according to the distance difference between the target obstacle and the interaction point after the target obstacle executes the driving strategy target duration; wherein the target duration is the duration for the unmanned equipment to pass through the interaction point normally; With the goal of minimizing the traffic difference and the spacing difference, determining a first control parameter used when the target obstacle moves toward the interaction point according to the driving strategy; Determining a first control parameter used when the target obstacle moves toward the interaction point according to the driving strategy specifically includes: If the driving strategy is a leading driving strategy, the preset control parameter is determined as the first control parameter used in the process of the target obstacle moving toward the interaction point according to the driving strategy.

2. The method according to claim 1, characterized in that Determine the target obstacles that interact with the unmanned equipment, including: Determine a planned path corresponding to the unmanned equipment and a predicted path corresponding to each obstacle within a preset range of the unmanned equipment; For each obstacle, determining whether there is an intersection between the planned path and the predicted path corresponding to the obstacle; According to the judgment result, a target obstacle that interacts with the unmanned equipment is determined.

3. The method according to claim 2, characterized in that According to the judgment result, the target obstacle with which the unmanned equipment interacts is determined, specifically including: If there is an intersection between the planned path and the predicted path corresponding to the obstacle, judging whether the obstacle is an obstacle that interacts with the unmanned device according to the time it takes for the unmanned device to pass through the intersection and the time it takes for the obstacle to pass through the intersection; If so, the obstacle is determined as a target obstacle that interacts with the unmanned device.

4. The method according to claim 1, characterized in that According to the motion state of the target obstacle at the current moment, predicting the driving strategy adopted by the target obstacle through the interaction point in consideration of the motion state of the unmanned equipment at the current moment, specifically includes: The predicted motion state of the target obstacle at the current moment predicted at the previous moment and the actual motion state of the target obstacle at the current moment are input into the prediction model, so as to determine the reward corresponding to each driving strategy according to the difference between the predicted motion state and the actual motion state through the prediction model, and predict the driving strategy adopted by the target obstacle to pass through the interaction point in consideration of the motion state of the unmanned equipment at the current moment according to the reward corresponding to each driving strategy.

5. The method according to claim 4, characterized in that Determining a reward corresponding to each driving strategy according to the difference between the predicted motion state and the actual motion state, specifically including: For each driving strategy, according to the difference between the predicted motion state corresponding to the driving strategy and the actual motion state, at least one reward indicator corresponding to the driving strategy is determined, and the at least one reward indicator includes: a safety reward indicator, a speed reward indicator, and a position reward indicator; A reward corresponding to the driving strategy is determined according to the at least one reward indicator.

6. The method according to claim 4, characterized in that According to the reward corresponding to each driving strategy, the driving strategy adopted by the target obstacle through the interaction point is predicted taking into account the current motion state of the unmanned device, specifically including: Determining the conditional probability of the target obstacle executing the real driving strategy according to the reward corresponding to each driving strategy, wherein the real driving strategy refers to the driving strategy adopted by the target obstacle to reach the real motion state; Determining the actual probability of the target obstacle executing each driving strategy at the current moment according to the conditional probability and the predetermined initial probability of the target obstacle executing each driving strategy; According to the actual probability of the target obstacle executing each driving strategy at the current moment, the driving strategy adopted by the target obstacle to pass through the interaction point is predicted while considering the motion state of the unmanned equipment at the current moment.

7. The method according to claim 6, characterized in that The method further comprises: According to the actual probability of the target obstacle executing each driving strategy at the current moment, the initial probability of the target obstacle executing each driving strategy is updated, and the updated probability is re-determined as the initial probability, so as to predict the driving strategy adopted by the target obstacle to pass through the interaction point in the future according to the re-determined initial probability.

8. The method according to claim 6, characterized in that Determining the initial probability of the target obstacle executing each driving strategy includes: When the target obstacle is initially determined to be an obstacle that interacts with the unmanned equipment, the initial probability of the target obstacle executing each driving strategy is determined based on the time it takes for the target obstacle to pass through the interaction point, the time it takes for the unmanned equipment to pass through the interaction point, and preset traffic rules.

9. An unmanned equipment control device, characterized in that: include: A first determination module is used to determine a target obstacle that interacts with the unmanned device, and predict a location point where the target obstacle interacts with the unmanned device as an interaction point; A prediction module, configured to predict, based on the motion state of the target obstacle at the current moment, a driving strategy adopted by the target obstacle to pass through the interaction point in consideration of the motion state of the unmanned equipment at the current moment; A second determination module is used to determine a first control parameter used in a process in which the target obstacle moves to the interaction point according to the driving strategy; A control module is used to determine a second control parameter for the unmanned device to pass through the interaction point according to the first control parameter, and control the unmanned device based on the second control parameter; the first control parameter used in the process of determining the target obstacle to go to the interaction point according to the driving strategy specifically includes: If the driving strategy is a yield driving strategy, determining a traffic difference corresponding to the target obstacle according to a difference between a time length for the target obstacle to pass through the interaction point when executing the driving strategy and a time length for the target obstacle to pass through the interaction point normally; Determine the spacing difference corresponding to the target obstacle according to the distance difference between the target obstacle and the interaction point after the target obstacle executes the driving strategy target duration; wherein the target duration is the duration for the unmanned equipment to pass through the interaction point normally; With the goal of minimizing the traffic difference and the spacing difference, determining a first control parameter used when the target obstacle moves toward the interaction point according to the driving strategy; Determining a first control parameter used when the target obstacle moves toward the interaction point according to the driving strategy specifically includes: If the driving strategy is a leading driving strategy, the preset control parameter is determined as the first control parameter used in the process of the target obstacle moving toward the interaction point according to the driving strategy.

10. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

11. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method described in any one of claims 1 to 8 is implemented.

Images