Automatic driving automobile battery replacing device transportation system

Abstract

The invention discloses an automatic driving automobile battery replacing device transportation system. Comprising a battery changing unit, a control unit and a communication unit. The battery replacing unit comprises a battery replacing vehicle, a battery replacing track, a high-definition camera and a new battery pack. The control unit comprises a user side and a central controller. A user side contacts with the central controller by sending real-time vehicle battery information, the central controller predicts the power-off time and the corresponding position of an electric vehicle at the user side and estimates the specific duration needed by the battery replacing vehicle to arrive at the electric vehicle, and an instruction is sent through a communication unit to enable the battery replacing vehicle to start according to a specified route. The purpose that the battery changing vehicle arrives when the electric vehicle is powered off is achieved, and a football team training optimization algorithm is adopted to optimize the route of the battery changing vehicle. Compared with the prior art, the method has the advantages of high reliability, strong autonomy, labor cost saving and the like, and the path of the battery changing vehicle is optimized through the football team training algorithm, so that the transportation efficiency and the successful transportation probability are improved.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent transportation technology, and specifically relates to an automatic driving transportation system for a vehicle battery replacement device. Background Art

[0002] New energy vehicles have been developed for many years, and their production and sales have been growing rapidly. However, the key technology of batteries has not made any substantial breakthroughs, and the anxiety of car owners about driving range has not been well resolved. It takes less than 5 minutes to refuel a fuel car, while electric cars generally take 6 to 8 hours to charge slowly, and at least 30 minutes to charge quickly. If there is a queue, it takes even longer, which seriously affects the car owner's car experience. In addition, fast charging will reduce the service life of the battery and increase the cost of using the car.

[0003] In order to solve industry pain points such as long charging time, short battery life, safety hazards, high investment costs and high site requirements of existing battery replacement methods, this application proposes a cost-controllable method for replacing battery packs anytime, anywhere, quickly and efficiently. Summary of the invention

[0004] Purpose of the invention: In response to the problems in the background technology, the present invention provides an autonomous driving vehicle battery replacement device transportation system, which can achieve cost-controlled, anytime, anywhere, fast and efficient replacement of battery packs.

[0005] Technical solution: The present invention discloses an autonomous driving vehicle battery replacement device transportation system, including a battery replacement unit, a control unit, and a communication unit. The battery replacement unit includes a battery replacement vehicle, a battery replacement track, a high-definition camera, and a new battery pack; the control unit includes a user terminal and a central controller;

[0006] When the battery reserve of the user-side electric vehicle is in a critical state, the current battery status and the expected driving route are automatically transmitted to the central controller through the communication unit. The central controller predicts the power-off time and corresponding position of the user-side electric vehicle based on the real-time data obtained, and estimates the specific time required for the battery-swapping vehicle to reach the electric vehicle. The communication unit sends instructions to let the battery-swapping vehicle set off along the designated route, so as to achieve the purpose of the battery-swapping vehicle arriving when the electric vehicle is powered off, and uses the football team training optimization algorithm to optimize the route of the battery-swapping vehicle. After the battery-swapping vehicle arrives, it uses a high-definition camera to shoot the vehicle and road conditions, and the central controller completes the battery-swapping work by controlling the battery-swapping vehicle's robotic arm.

[0007] Furthermore, the objective function is designed, and the following objective function is constructed with the minimum search and rescue time as the goal:

[0008]

[0009]

[0010] st is a constraint condition. Constraint (1) indicates that the power consumed by the battery swap vehicle during driving cannot exceed the maximum battery power; constraint (2) ensures that the transportation task must be completed within the scheduled time; constraint (3) prevents the battery swap vehicle from entering the threat area; where C is the search and rescue time; is the weight parameter, d i is the distance from the i-1th point to the i-th point of the electric car, n is the number of nodes in the path, Indicates the turning angle of the electric vehicle, N a represents the radius of each turn of the battery-swapping vehicle, x is the number of turns, q represents the set of all obstacles, and all obstacles are considered as cylinders, D represents the diameter of the battery-swapping vehicle, and R q represents the obstacle radius, P X is the battery power.

[0011] Furthermore, the football team training algorithm is used to optimize the path and improve the transportation efficiency of the car-swap electricity. The implementation process of the football team training optimization algorithm is as follows:

[0012] Step 1: Optimize and initialize the players, set the number of players to N; each player is randomly placed in the space as a candidate solution x = (x1, ..., xi) represented by a three-dimensional vector, set the maximum number of iterations of the algorithm, and set the search space of the algorithm according to the number of iterations;

[0013] Step 2: The initial population is brought into the objective function for evaluation, so that a set of corresponding group values ​​of the shortest path length and the minimum threat area are generated for each player;

[0014] Step 3: Football training sessions are divided into three parts: collective training, small group training and individual additional training;

[0015] 1) At the beginning of collective training, players are divided into four different types: Followers, Explorers, Contemplators, and Fluctuators. In each iteration, players will randomly change their type;

[0016] Followers randomly approach the top player's position in different ways, using the following equation:

[0017]

[0018] where the current best player is defined as k is the number of iterations, is the value on dimension j; the current player is defined as i is the number of players, is the value on dimension j, is the state of the player in dimension j after training, is the state of the player in dimension j before training, unif is a random number uniformly distributed in the range [0,1), and the optimized followers Represents the objective function d i The value in , i.e., the distance from the i-1th point to the i-th point of the battery-changing vehicle, indicates that the path at this time is the current optimal path;

[0019] The explorer adopts a strategy to minimize the risk of becoming the worst performer, as follows:

[0020]

[0021] In the formula, unif 1 ,unif 2 represents a random number uniformly distributed in the range [0, 1). The current worst player is defined as k is the number of iterations, is the value in dimension j, is the state of the player on dimension j after training. At this time, the explorer is an optimized follower, representing d in the objective function. i The value in represents the better path found after the algorithm iteration at this time;

[0022] Meditators identify and close gaps in each dimension using the following equation:

[0023]

[0024] In dimension j, is the difference vector between the current best player and the worst player. The meditator at this time represents the optimized explorer and represents d in the objective function. i The value in indicates that the path is further optimized at this time;

[0025] The change of the fluctuator player state is defined as follows:

[0026]

[0027] Among them, t(k) is a random number with t distribution, and the degree of freedom is the current number of iterations. As the degree of freedom increases, the probability of t distribution approaching the middle value 0 becomes higher and higher, and the distribution at both ends gradually decreases, and will become closer and closer to the normal distribution; as the number of iterations increases, the degree of fluctuation will become smaller and smaller, and gradually change from global search to local search. The volatile is used as a reference variable to compare with the optimized meditator, and the algorithm will select a better value to replace the d of the objective function. i ;

[0028] 2) After the collective training, the football training process has entered the group training stage. The group training is defined as three states: optimal learning, random learning and random communication. The learning probability is defined as L study , the communication probability is L comm , the player will randomly choose the state in each iteration;

[0029] Under the optimal learning strategy, players have a certain probability of directly adopting the ability values ​​of the top players in the team as learning goals in each dimension to optimize and improve their own abilities. The formula is defined as follows

[0030]

[0031] Where, L study represents the learning probability, Is the best player in group l, team l Represents group l, is the state of the player in the j dimension after optimal learning, is the state of the player in the j dimension before optimal learning, and after optimal learning Represents N in the objective function a , which indicates the optimal turning radius found by the algorithm at this time;

[0032] Under the random learning strategy, players select the ability value of any player in the team for direct learning in each dimension according to a certain probability. The formula is defined as follows:

[0033]

[0034] In the formula, is a random player of dimension j in group l, is the state of the player in the j dimension after random learning, is the state of the player in the j dimension before random learning, and the value after random learning represents the value in the objective function Indicates the optimal turning angle found by the algorithm at this time;

[0035] In the random communication strategy, in each dimension, the player will exchange knowledge and experience with any player in the group according to a specific probability. The formula is defined as follows:

[0036]

[0037] Among them, stoch is a normally distributed random number, is the state of the player in the j dimension before random communication, is the state of the player in the j dimension after random communication, is the state of the random player of dimension j in group 1 after random communication, is the state of the j-dimensional random player in group 1 before random communication. Multiplying it by (1+stoch) represents the understanding of the other's ability by the two players. Random communication will be used as a variable, and the value generated will be compared with the value of optimal learning and random learning. The algorithm will select the better value to replace the objective function.

[0038] 3) In the individual strengthening training mechanism, better physical fitness indicators are used to replace the original lower-level indicators. After the update process is completed, the coach will select the most outstanding players for special training. The training formula is as follows:

[0039]

[0040] Where Gauss is the standard deviation of the Gaussian distribution in the Gaussian variation, Cauchy is the scale parameter of the Cauchy distribution in the Cauchy variation, It is the state of the player before individual enhancement. is the state of the player after individual reinforcement. Cauchy and Gaussian joint variation is used to describe the individual's additional training. Represents D+R in the objective function q -d q , indicating that the path at this time is the optimal path after the algorithm has searched for the best path;

[0041] 4) Determine whether the termination condition is met. If the termination condition is not met, return to step 2); if the termination condition is met, go to step 4);

[0042] 5) Output the optimal objective function value.

[0043] Furthermore, the path optimization of the football team training algorithm is improved by introducing an adaptive weight factor so that the players can update their positions using a smaller adaptive weight. The improved formula is as follows:

[0044]

[0045] In the formula, is the new position of the particle after updating after the improved algorithm; ω is the adaptive weight factor.

[0046] Furthermore, when the battery-swapping vehicle arrives at the electric vehicle, the following steps are adopted to swap the battery: the first step is preliminary preparation, installing a simple ramp, the second step is the electric vehicle driving up the ramp and removing the middle section, the third step is the battery-swapping vehicle entering the bottom of the battery and the mechanical arm rising and supporting the battery pack, the fourth step is removing the battery pack, the fifth step is the battery-swapping vehicle descending with the empty battery pack and leaving the vehicle, the sixth step is the battery-swapping vehicle carrying the fully charged battery pack into the bottom of the vehicle and lifting it up with a mechanical arm, the seventh step is installing the battery pack, the eighth step is the battery-swapping vehicle evacuating, and the ninth step is the vehicle reversing and leaving the battery-swapping area.

[0047] Beneficial effects:

[0048] 1. The path planning in the present invention is one of the important components of the battery-swapping vehicle and is a key technology for realizing the automatic driving of the battery-swapping vehicle. A reasonable delivery path will not only improve the driving efficiency of the battery-swapping vehicle in complex environments and significantly reduce the time cost, but also effectively increase the probability of successful transportation. It can alleviate the mileage anxiety of electric vehicle owners to a certain extent, which is conducive to improving the overall satisfaction and trust of users in electric vehicles, thereby helping to accelerate the popularization of electric vehicles, promote the healthy development of the new energy vehicle industry, drive the development of related industrial chains, and promote the creation of new employment opportunities. Through the calculation of the central controller, the battery-swapping vehicle can arrive when the electric vehicle is powered off.

[0049] 2. The present invention adopts unmanned driving technology, which is easy to operate and does not require professional personnel to operate. It can respond quickly in complex traffic environments and significantly reduce traffic accidents caused by human errors, thereby greatly improving road safety. It reduces energy consumption by intelligently planning driving routes and optimizing driving behaviors, which helps to reduce dependence on fossil energy, promote the transformation of energy consumption to green and low-carbon, improve the efficiency and quality of transportation services, and reduce labor costs.

[0050] 3. The present invention takes the shortest search and rescue time as the objective function, uses the football team training algorithm to optimize the transportation path, and through the excellent optimization mechanism of the football team training algorithm, continuously iterates through followers, explorers, contemplators and undulators to obtain the optimal d in the objective function. i , through optimal training, random training, random communication training to obtain the optimal objective function Obtain the optimal D+R in the objective function through individual reinforcement training q -d q , so that we can obtain the shortest transportation time. Through the experimental results, we can know that the football team training algorithm has excellent performance and converges very quickly. The football team training algorithm can respond to changes in the transportation process in real time, quickly adjust transportation strategies, and improve the flexibility and adaptability of logistics scheduling. By accurately calculating and optimizing transportation routes, it can effectively reduce fuel consumption, reduce vehicle wear and maintenance costs, thereby achieving a substantial reduction in logistics costs. Compared with other transportation algorithms, it has advantages in improving the transportation efficiency and successful transportation probability of battery-swapping vehicles in complex environments, and introduces adaptive weight factors so that players use smaller adaptive weights to update their positions, thereby improving the local optimization ability of the football team training algorithm.

[0051] 4. In the individual reinforcement training mechanism of the present invention, Cauchy and Gaussian joint variation is used to describe individual additional training. The reason for choosing Gaussian Cauchy distribution is that in the early stages of training, everyone's level is usually not high, so the best players have a greater probability of getting a greater improvement. At this time, the Cauchy distribution function occupies a large proportion, which can effectively provide a large range of improvement for players, which is conducive to global search. As the number of iterations increases, it becomes more and more difficult to improve the player's ability. At this time, the Gaussian distribution accounts for a large proportion, and the player's promotion range gradually decreases, which is more conducive to local search. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a structural schematic diagram of the present invention;

[0053] Figure 2 The battery replacement step of the present invention

[0054] Figure 3 This is a simulation diagram of the football team training optimization algorithm proposed by the present invention. DETAILED DESCRIPTION

[0055] The present invention will be further described in detail below in conjunction with the accompanying drawings.

[0056] like Figure 1 As shown, the autonomous driving vehicle battery replacement device transportation system disclosed in the present invention includes a battery replacement unit, a control unit and a communication unit. The battery replacement unit includes a battery replacement vehicle, a battery replacement track, a high-definition camera and a new battery pack. The control unit includes a user terminal and a central controller. The communication unit includes a wireless communication terminal, a data message and an intelligent gateway.

[0057] When the battery swap vehicle receives the transportation task, it uses the football team training optimization algorithm to optimize the path, improve the transportation efficiency of the battery swap vehicle in complex environments and save transportation costs. After arriving at the designated location, the following steps are used to swap the battery. The first step is preliminary preparation, installing a simple ramp, the second step is for the vehicle to drive up the ramp and remove the middle section, the third step is for the battery swap vehicle to enter the bottom of the battery and the mechanical arm rises and supports the battery pack, the fourth step is to remove the battery pack, the fifth step is for the battery swap vehicle to descend with the empty battery pack and leave the vehicle, the sixth step is for the battery swap vehicle to carry the fully charged battery pack to the bottom of the vehicle and lift it up with a mechanical arm, the seventh step is to install the battery pack, the eighth step is for the battery swap vehicle to evacuate, and the ninth step is for the vehicle to reverse and drive out of the battery swap area.

[0058] The data message in the communication unit contains the surrounding environment and the operating status of the battery-swapping vehicle itself; the intelligent gateway filters and screens the content of the data message to ensure data transmission between the chip and the central controller.

[0059] The present invention utilizes the football team training algorithm to optimize the path, and the implementation process is as follows:

[0060] 1) Optimize the initialization of the players. Set the number of players to N; each player is randomly placed in the space as a candidate solution x = (x1, ..., xi) represented by a three-dimensional vector, set the maximum number of iterations of the algorithm, and set the search space of the algorithm according to the number of iterations.

[0061] 2) The initial population is brought into the objective function for evaluation. Each player is made to generate a corresponding group value of the shortest path length and the minimum threat area. The objective function formula is as follows:

[0062]

[0063] Constraint (1) indicates that the power consumed by the battery swap vehicle during driving cannot exceed the maximum battery power; constraint (2) ensures that the transportation task must be completed within the scheduled time; constraint (3) prevents the battery swap vehicle from entering the threat area.

[0064] In the formula, C is the search and rescue time; st is the constraint condition; is the weight parameter, d i is the distance from the i-1th point to the i-th point of the switching vehicle, and n is the number of nodes in the path. Indicates the turning angle of the electric vehicle, N a represents the radius of each turn of the battery-swapping vehicle, x is the number of turns, q represents the set of all obstacles, and all obstacles are considered as cylinders, D represents the diameter of the battery-swapping vehicle, and R q Represents the obstacle radius. X is the battery power.

[0065] 3) The football team training algorithm is implemented by simulating the behavior of players in high-level football training sessions. Football training sessions are divided into three parts: collective training, group training, and individual intensive training.

[0066] 3.1) At the beginning of collective training, we divided the players into four different types: Followers, Explorers, Contemplators, and Fluctuators. In each iteration, the players will randomly change their type.

[0067] Followers are enthusiastic imitators of the current top players. They strive to imitate the style of top players in all aspects and are eager to reach their level. However, due to their own limitations, they can usually only randomly approach the position of top players in different aspects. The formula is as follows:

[0068]

[0069] where the current best player is defined as k is the number of iterations, is its value on dimension j; the current player is defined as i is the number of players, is its value in dimension j, is the state of the player in dimension j after training, and unif is a random number uniformly distributed in the range [0,1). Represents the objective function d i The value in indicates that the path at this time is the current optimal path.

[0070] Explorers show higher rational thinking than followers. In the evaluation process, they not only pay attention to the performance of top players, but also observe the players who perform poorly. Therefore, while pursuing excellence, they also actively adopt strategies to minimize the risk of becoming the worst performer. The equation is as follows:

[0071]

[0072] In the formula, the current worst player is defined as is its value in dimension j. is the state of the player on dimension j after training. The explorer at this time is an optimized follower, representing d in the objective function i The value in represents the better path found after the algorithm iteration.

[0073] The meditators show a higher level of awareness than the individuals before them, and they directly see the gap between top players and underperforming players. They work to identify and close the gaps in various dimensions in order to achieve overall improvement in their abilities. The equation is as follows:

[0074]

[0075] In dimension j, is the difference vector between the current best player and the worst player. The meditator at this time represents the optimized explorer and represents d in the objective function. i The value in indicates that the path is further optimized at this time.

[0076] Fluctuators tend to train autonomously rather than imitate others. Due to their independent learning paths, their states may show a certain degree of volatility. However, as the training frequency increases (the number of iterations increases), the instability of these individual states will gradually decrease and tend to be stable. We define the change of player states as follows:

[0077]

[0078] Among them, t(k) is a random number with t distribution, and its degree of freedom is the current number of iterations. As the degree of freedom increases, the probability of t distribution approaching the middle value (0) becomes higher and higher, and the distribution at both ends gradually decreases, and will become closer and closer to the normal distribution. Therefore, as the number of iterations increases, the degree of fluctuation will become smaller and smaller, and gradually change from global search to local search. The volatile will be used as a reference variable to compare with the optimized meditator, and the algorithm will select a better value to replace the d of the objective function. i .

[0079] 3.2) After the collective training, the football training process has entered the group training stage. We define group training as three states: optimal learning, random learning, and random communication. We define the learning probability as L study , the communication probability is L comm , the player will randomly choose the state in each iteration.

[0080] Under the optimal learning strategy, players have a certain probability of directly adopting the ability values ​​of the top players in the team as learning goals in all dimensions to optimize and improve their own abilities. The formula is defined as follows

[0081]

[0082] In the formula, Is the best player in group l, team l Represents group l, is the state of the player in the j dimension after optimal learning. Represents N in the objective function a , which indicates the optimal turning radius found by the algorithm at this time.

[0083] Under the random learning strategy, players select the ability value of any player in the team for direct learning in each dimension according to a certain probability, as a means of improving ability. The formula is defined as follows:

[0084]

[0085] In the formula, is a random player in group l, is the j-dimension of a random player in group l, is the state of the player in the j dimension after random learning. The value after random learning represents the value in the objective function Indicates the optimal turning angle found by the algorithm at this time.

[0086] In the random communication strategy, although the learning process in the training process is an important part, the interaction between players has a more significant impact on the improvement of ability. In each dimension, the player will exchange knowledge and experience with any player in the group according to a specific probability to promote the comprehensive improvement of ability. The formula is defined as follows:

[0087]

[0088] in, is a random player in group l, j is the j-th dimension of a random player in group l, and the two players exchanged their abilities in the j-th dimension. is the state of the player in the j dimension before random communication, is the state of the player in the j dimension after random communication, is the state of the random player of dimension j in group 1 after random communication, is the state of the j-dimensional random player in group 1 before random communication. Stoch is a normally distributed random number. Multiplying (1+stoch) represents the understanding of the two players on each other's abilities. Random communication will be used as a variable, and the value generated will be compared with the values ​​of optimal learning and random learning. The algorithm will select the better value to replace the objective function.

[0089] 3.3) In the individual intensive training mechanism, after the group training phase, the players' new physical indicators must be re-evaluated. At this time, better physical indicators should be used to replace the original lower-level indicators to achieve the optimization and update of the players' status. After the update process is completed, the coach will select the most outstanding players for special training, aiming to further improve their abilities so that they can more effectively lead and motivate the training status of other players. The training formula is as follows:

[0090]

[0091] Where Gauss is the standard deviation of the Gaussian distribution in the Gaussian variation, Cauchy is the scale parameter of the Cauchy distribution in the Cauchy variation, It is the state of the player before individual enhancement. is the state of the player after individual reinforcement, and Cauchy and Gaussian joint variation is used to describe individual additional training k is the number of iterations. The reason for choosing Gaussian Cauchy distribution is that in the early stages of training, everyone’s level is usually not high, so the best players have a greater probability of getting a bigger improvement. At this time, the Cauchy distribution function occupies a large proportion, which can effectively provide a large range of improvements for players, which is conducive to global search. As the number of iterations increases, it becomes more and more difficult to improve the player’s ability. At this time, the Gaussian distribution accounts for a large proportion, and the player’s promotion range gradually decreases, which is more conducive to local search. After individual reinforcement Represents D+R in the objective function q -d q , indicating that the path at this time is the optimal path after the algorithm has searched for the optimal path.

[0092] 4) Determine whether the termination condition is met. If the termination condition is not met, return to step 3); if the termination condition is met, go to step 5).

[0093] 5) Output the optimal objective function value.

[0094] At the same time, the football team training algorithm is improved by introducing an adaptive weight factor in step 3) so that the players use a smaller adaptive weight to update their positions, thereby improving the local optimization ability of the football team training algorithm. The improved formula is as follows:

[0095]

[0096] In the formula, is the new position of the particle after updating after the improved algorithm; ω is the adaptive weight factor.

[0097] Figure 3 The simulation diagram of the football team training optimization algorithm proposed by the present invention, the experimental results clearly show the change of the best fitness with the number of iterations and the 3D visualization of the parameter space. The chart on the left is the fitness curve, which shows the change of the best fitness of the football training team algorithm in the iteration process. The fitness value gradually decreases, indicating that the algorithm is looking for a better solution. The chart on the right is the parameter space, and the surface gradually decreases from yellow to blue, indicating the change of the function value, and the lowest point is in the center, indicating the position of the global optimal solution. From the experimental results, we can see that the football training team algorithm can find a better solution in the iteration process, and the fitness value gradually decreases, indicating that the algorithm has a strong global search ability. The curve tends to be stable in the later stage of iteration, indicating that the algorithm can converge to a better solution and avoid falling into the local optimum.

[0098] The above embodiments are only for illustrating the technical concept and features of the present invention, and their purpose is to enable people familiar with the technology to understand the content of the present invention and implement it accordingly, and they cannot be used to limit the protection scope of the present invention. Any equivalent transformation or modification made according to the spirit of the present invention should be included in the protection scope of the present invention.

Claims

1. An automatic driving vehicle battery replacement device transportation system, characterized in that: It includes a battery replacement unit, a control unit, and a communication unit. The battery replacement unit includes a battery replacement vehicle, a battery replacement track, a high-definition camera, and a new battery pack. The control unit includes a user terminal and a central controller. When the battery reserve of the user-side electric vehicle is in a critical state, the current battery status and the expected driving route are automatically transmitted to the central controller through the communication unit. The central controller predicts the power-off time and corresponding position of the user-side electric vehicle based on the real-time data obtained, and estimates the specific time required for the battery-swapping vehicle to reach the electric vehicle. The communication unit sends instructions to let the battery-swapping vehicle set off along the designated route, so as to achieve the purpose of the battery-swapping vehicle arriving when the electric vehicle is powered off, and uses the football team training optimization algorithm to optimize the route of the battery-swapping vehicle. After the battery-swapping vehicle arrives, it uses a high-definition camera to shoot the vehicle and road conditions, and the central controller completes the battery-swapping work by controlling the battery-swapping vehicle's robotic arm.

2. The automatic driving vehicle battery replacement device transportation system according to claim 1 is characterized in that: Design the objective function and construct the following objective function with the minimum search and rescue time as the goal: st is a constraint condition. Constraint (1) indicates that the power consumed by the battery swap vehicle during driving cannot exceed the maximum battery power; constraint (2) ensures that the transportation task must be completed within the scheduled time; constraint (3) prevents the battery swap vehicle from entering the threat area; where C is the search and rescue time; is the weight parameter, d i is the distance from the i-1th point to the i-th point of the electric vehicle, n is the number of nodes in the path, θ a Indicates the turning angle of the electric vehicle, N a represents the radius of each turn of the battery-swapping vehicle, x is the number of turns, q represents the set of all obstacles, and all obstacles are considered as cylinders, D represents the diameter of the battery-swapping vehicle, and R q represents the obstacle radius, P X is the battery power.

3. The automatic driving vehicle battery replacement device transportation system according to claim 2 is characterized in that: Use the football team training algorithm to optimize the path and improve the transportation efficiency of the car-swap battery. The implementation process of the football team training optimization algorithm is as follows: Step 1: Optimize and initialize the players, set the number of players to N; each player is randomly placed in the space as a candidate solution x = (x1, ..., xi) represented by a three-dimensional vector, set the maximum number of iterations of the algorithm, and set the search space of the algorithm according to the number of iterations; Step 2: The initial population is brought into the objective function for evaluation, so that a set of corresponding group values ​​of the shortest path length and the minimum threat area are generated for each player; Step 3: Football training sessions are divided into three parts: collective training, small group training and individual additional training; 1) At the beginning of collective training, players are divided into four different types: Followers, Explorers, Contemplators, and Fluctuators. In each iteration, players will randomly change their type; Followers randomly approach the top player's position in different ways, using the following equation: where the current best player is defined as k is the number of iterations, is the value on dimension j; the current player is defined as i is the number of players, is the value on dimension j, is the state of the player in dimension j after training, is the state of the player in dimension j before training, unif is a random number uniformly distributed in the range [0,1), and the optimized followers Represents the objective function d i The value in , i.e., the distance from the i-1th point to the i-th point of the battery-changing vehicle, indicates that the path at this time is the current optimal path; The explorer adopts a strategy to minimize the risk of becoming the worst performer, as follows: Where unif1 and unif2 both represent a random number uniformly distributed in the range [0, 1). The current worst player is defined as k is the number of iterations, is the value in dimension j, is the state of the player on dimension j after training. At this time, the explorer is an optimized follower, representing d in the objective function. i The value in represents the better path found after the algorithm iteration at this time; Meditators identify and close gaps in each dimension using the following equation: In dimension j, is the difference vector between the current best player and the worst player. The meditator at this time represents the optimized explorer and represents d in the objective function. i The value in indicates that the path is further optimized at this time; The change of the fluctuator player state is defined as follows: Among them, t(k) is a random number with t distribution, and the degree of freedom is the current number of iterations. As the degree of freedom increases, the probability of t distribution approaching the middle value 0 becomes higher and higher, and the distribution at both ends gradually decreases, and will become closer and closer to the normal distribution; as the number of iterations increases, the degree of fluctuation will become smaller and smaller, and gradually change from global search to local search. The volatile is used as a reference variable to compare with the optimized meditator, and the algorithm will select a better value to replace the d of the objective function. i ; 2) After the collective training, the football training process has entered the group training stage. The group training is defined as three states: optimal learning, random learning and random communication. The learning probability is defined as L study , the communication probability is L comm , the player will randomly choose the state in each iteration; Under the optimal learning strategy, players have a certain probability of directly adopting the ability values ​​of the top players in the team as learning goals in each dimension to optimize and improve their own abilities. The formula is defined as follows Where, L study represents the learning probability, Is the best player in group l, team l Represents group l, is the state of the player in the j dimension after optimal learning, is the state of the player in the j dimension before optimal learning, and after optimal learning Represents N in the objective function a , which indicates the optimal turning radius found by the algorithm at this time; Under the random learning strategy, players select the ability value of any player in the team for direct learning in each dimension according to a certain probability. The formula is defined as follows: In the formula, is a random player of dimension j in group l, is the state of the player in the j dimension after random learning, is the state of the player in the j dimension before random learning, and the value after random learning represents the value in the objective function Indicates the optimal turning angle found by the algorithm at this time; In the random communication strategy, in each dimension, the player will exchange knowledge and experience with any player in the group according to a specific probability. The formula is defined as follows: Among them, stoch is a normally distributed random number, L comm is the communication probability, is the state of the player in the j dimension before random communication, is the state of the player in the j dimension after random communication, is the state of the random player of dimension j in group 1 after random communication, is the state of the j-dimensional random player in group 1 before random communication. Multiplying it by (1+stoch) represents the understanding of the other's ability by the two players. Random communication will be used as a variable, and the value generated will be compared with the value of optimal learning and random learning. The algorithm will select the better value to replace the objective function. 3) In the individual strengthening training mechanism, better physical fitness indicators are used to replace the original lower-level indicators. After the update process is completed, the coach will select the most outstanding players for special training. The training formula is as follows: Where Gauss is the standard deviation of the Gaussian distribution in the Gaussian variation, Cauchy is the scale parameter of the Cauchy distribution in the Cauchy variation, It is the state of the player before individual enhancement. is the state of the player after individual reinforcement. Cauchy and Gaussian joint variation is used to describe the individual's additional training. Represents D+R in the objective function q -d q , indicating that the path at this time is the optimal path after the algorithm has searched for the best path; 4) Determine whether the termination condition is met. If the termination condition is not met, return to step 2); if the termination condition is met, go to step 4); 5) Output the optimal objective function value.

4. The automatic driving vehicle battery replacement device transportation system according to claim 3 is characterized in that: The path optimization of the football team training algorithm is improved by introducing an adaptive weight factor so that the players can update their positions using a smaller adaptive weight. The improved formula is as follows: In the formula, is the new position of the particle after updating after the improved algorithm; ω is the adaptive weight factor.

5. The automatic driving vehicle battery replacement device transportation system according to claim 1 is characterized in that: When the battery-swapping vehicle arrives at the electric vehicle, the following steps are used to swap the battery: the first step is preliminary preparation, installing a simple ramp, the second step is the electric car driving up the ramp and removing the middle section, the third step is the battery-swapping vehicle entering the bottom of the battery and the robotic arm rising and supporting the battery pack, the fourth step is removing the battery pack, the fifth step is the battery-swapping vehicle descending with the empty battery pack and leaving the vehicle, the sixth step is the battery-swapping vehicle carrying the fully charged battery pack into the bottom of the vehicle and lifting it up with a robotic arm, the seventh step is installing the battery pack, the eighth step is the battery-swapping vehicle evacuating, and the ninth step is the vehicle reversing to leave the battery-swapping area.

Images