An AI-based path planning method for automated downhole driving

Abstract

This invention provides an artificial intelligence-based method for underground autonomous driving path planning, relating to the field of autonomous driving control technology. It includes determining safe distances between vehicles; planning multiple transport lanes that only allow one-way transport by autonomous vehicles, with these lanes arranged repeatedly in a pattern of two single-lane lanes in the first direction and two single-lane lanes in the second direction; when a transport lane experiences a malfunction and vehicles cannot pass, vehicles facing the entrance of an evacuation lane sequentially enter the nearest evacuation lane; at least one transport lane on either side of the malfunctioning transport lane is designated as an evacuation lane, and vehicles facing the entrance of the evacuation lane adjust their distance to twice the safe distance, allowing vehicles exiting the evacuation lane to insert themselves between vehicles in the evacuation lane. This invention can efficiently evacuate congested autonomous vehicles and avoid congestion or collisions during the evacuation process.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving control technology, and in particular to an artificial intelligence-based method for downhole autonomous driving path planning. Background Technology

[0002] Mining using autonomous vehicles underground is a new, efficient, and safe method. Autonomous vehicles eliminate the need for human labor, are low-cost, highly efficient, and avoid personnel safety issues.

[0003] Currently, the working mode of autonomous vehicles is relatively simple. When a fault occurs in the underground roadway, preventing the autonomous vehicles from passing, the efficiency of evacuating the congested autonomous vehicles is low. Furthermore, the evacuation process is prone to problems such as congestion or collisions of autonomous vehicles, leading to a decrease in the efficiency of automated transportation operations in coal mines. Summary of the Invention

[0004] This invention provides an artificial intelligence-based path planning method for automated underground mining operations to address the problems mentioned in the background section. The technical solution of this invention is as follows:

[0005] An artificial intelligence-based method for downhole autonomous driving path planning includes:

[0006] Determine the safe distance between autonomous vehicles, where the distance between different autonomous vehicles is the safe distance;

[0007] Multiple transport lanes are planned that allow only one-way transport by autonomous vehicles; wherein, the multiple transport lanes are arranged repeatedly in the form of two first-direction one-way and two second-direction one-way, so that one of the transport lanes on both sides of each transport lane is the first direction and the other is the second direction, the first direction is the ore entry direction and the second direction is the ore exit direction, and each transport lane is connected to the transport lanes on both sides of it, which are limited to different driving directions, through at least one evacuation passage;

[0008] When a fault occurs in one of the transport lanes, preventing the autonomous vehicles from passing, all the blocked autonomous vehicles in that transport lane stop moving. The autonomous vehicles facing the entrance of the evacuation channel enter the nearest evacuation channel in sequence. If there is no evacuation channel in the direction the vehicles are facing, after all the autonomous vehicles facing the entrance of the evacuation channel have entered the evacuation channel, they reverse to the entrance of the evacuation channel and enter in sequence. The timer starts after the first autonomous vehicle enters the evacuation channel. The next autonomous vehicle drives to the entrance of the evacuation channel and waits for the preset time to elapse before the next autonomous vehicle enters the evacuation channel.

[0009] When a fault occurs in one of the transport lanes, preventing the autonomous vehicles from passing, at least one of the transport lanes on both sides of the fault lane becomes an evacuation lane. In the evacuation lane, the autonomous vehicles facing the entrance of the evacuation channel adjust their distance to twice the safe distance, so that the autonomous vehicles exiting the evacuation channel insert themselves between the autonomous vehicles in the evacuation lane.

[0010] Optionally, the autonomous driving vehicle facing the entrance of the evacuation route adjusts its distance to twice the safe distance, which is achieved in the following way:

[0011] S1. Based on the distance and preset speed in the evacuation channel, determine the first acceleration of the autonomous vehicle that is moving in uniform acceleration in the evacuation channel. The autonomous vehicle at the entrance of the evacuation channel starts to move in uniform acceleration with the first acceleration, so that after the evacuation time, it reaches the end of the evacuation channel at a preset speed. The preset speed is the driving speed of the autonomous vehicle in the evacuation channel.

[0012] S2, when an autonomous vehicle enters the evacuation passage and begins to accelerate at the first acceleration, prediction begins. After predicting the evacuation time, the autonomous vehicle in the evacuation passage that has not passed the preset location and is closest to the preset location is marked as the target vehicle; wherein, the preset location is the position in the evacuation passage along the oncoming vehicle direction at the safe distance from the entrance of the evacuation passage.

[0013] S3, the target vehicle determines a second acceleration and a third acceleration based on its current position and the distance from the preset location. During the evacuation time, the target vehicle adjusts its speed by first decelerating and then accelerating with the second acceleration and the third acceleration in sequence, so that the target vehicle reaches the preset location at the preset speed.

[0014] S4, after the preceding autonomous vehicle enters the evacuation channel, the automated vehicle in the transport lane where the fault occurred waits at the entrance of the evacuation channel. After the preceding autonomous vehicle exits the evacuation channel, it begins to accelerate with an initial velocity of 0 and the first acceleration, repeating S2-S3.

[0015] Optionally, the absolute values ​​of the second acceleration and the third acceleration are equal, their directions are opposite, and the durations of the second acceleration and the third acceleration are the same, both being half of the evacuation time.

[0016] Optionally, the absolute values ​​of both the second acceleration and the third acceleration are 4(v0T-S) / T. 2Where v0 is the preset speed, T is the evacuation time, and S is the distance between the target vehicle and the preset location.

[0017] Optionally, an AI-based method for downhole autonomous driving path planning also includes:

[0018] A constraint is set whereby, when the distance between any autonomous vehicle and the vehicle in front is less than the minimum distance, the autonomous vehicle activates adaptive cruise control to ensure that the autonomous vehicle maintains at least the minimum distance from the vehicle in front.

[0019] Optionally, the adaptive cruise is implemented using edge computing.

[0020] Optionally, the evacuation passage between two transport lanes with opposite directions of travel includes, in sequence, an arc-shaped section and a straight section along the vehicle-defined travel direction.

[0021] Compared with the prior art, the present invention has at least the following beneficial effects:

[0022] In this invention, multiple transport lanes are planned and set up, each designed as a one-way street, preventing congestion caused by passing and U-turns, and greatly ensuring transport efficiency. Furthermore, using two transport lanes in the first direction and two in the second direction as basic units, multiple lanes are arranged in a cyclical pattern, with the traffic directions of the transport lanes on both sides of each transport lane being opposite. With this arrangement, if any transport lane experiences a malfunction, such as a collapse or blockage, there is always a transport lane with the same traffic direction as the affected lane that serves as an evacuation lane for the automated driving vehicles to evacuate. During the evacuation process, the autonomous vehicles traveling in the transport lanes, which serve as evacuation routes, are already maintaining a safe distance. In order to both continue the original transport tasks of the autonomous vehicles in the evacuation routes to a certain extent and accommodate the autonomous vehicles in the transport lanes that are blocked by malfunctions, when the evacuation task is issued upon detecting a malfunction, the vehicles in the evacuation routes adjust their distance to twice the safe distance. When the autonomous vehicles in the evacuation route drive out, they are precisely at the preset speed and inserted between the two autonomous vehicles that have completed the distance adjustment, ensuring the safe driving of the autonomous vehicles.

[0023] In summary, when an autonomous vehicle is unable to pass through an underground roadway due to a fault or even becomes blocked, this invention can efficiently evacuate the blocked autonomous vehicle and avoid problems such as congestion or collisions during the evacuation process, thereby improving the efficiency of automated transportation operations in coal mines. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is a schematic diagram of a transportation roadway and evacuation passage in an artificial intelligence-based underground autonomous driving path planning method provided in an embodiment of the present invention.

[0026] In the picture:

[0027] 1-Transport roadway;

[0028] 2-Evacuation routes. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0030] like Figure 1 As shown in the figure, an embodiment of the present invention provides an artificial intelligence-based path planning method for downhole autonomous driving, comprising:

[0031] Determine the safe distance between autonomous vehicles; the distance between different autonomous vehicles is the safe distance.

[0032] The plan includes multiple transport tunnels that allow only one-way transport by autonomous vehicles. These transport tunnels are arranged in a repeating pattern of two first-direction one-way tunnels and two second-direction one-way tunnels, so that each transport tunnel has one first-direction tunnel on each side and the other second-direction tunnel on each side. The first-direction tunnel is the direction of entering the mine, and the second-direction tunnel is the direction of exiting the mine. Each transport tunnel is connected to the transport tunnels on both sides of it, which are restricted to different directions of travel, through at least one evacuation passage.

[0033] When a fault occurs in a transport lane, preventing autonomous vehicles from passing, all autonomous vehicles blocked in that transport lane stop moving. Autonomous vehicles facing the entrance of the evacuation channel enter the nearest evacuation channel in sequence. If there is no evacuation channel in the direction the vehicles are facing, they reverse to the entrance of the evacuation channel and enter in sequence after all the autonomous vehicles facing the entrance of the evacuation channel have entered the evacuation channel. The timer starts after the first autonomous vehicle enters the evacuation channel, and the next autonomous vehicle drives to the entrance of the evacuation channel and takes its position. After the preset time has elapsed, the next autonomous vehicle enters the evacuation channel.

[0034] When a fault occurs in a transport lane, preventing autonomous vehicles from passing, at least one of the transport lanes on both sides of the fault becomes an evacuation lane. In the evacuation lane, the autonomous vehicles facing the entrance of the evacuation channel adjust their distance to twice the safe distance, so that the autonomous vehicles exiting the evacuation channel insert themselves between the autonomous vehicles in the evacuation lane.

[0035] In this embodiment, multiple transport lanes are planned and set up, each designed as a one-way street to prevent congestion caused by passing and U-turns, greatly ensuring transport efficiency. Furthermore, using two transport lanes in the first direction and two in the second direction as basic units, multiple transport lanes are arranged in a cyclical pattern, with the travel directions of the transport lanes on both sides of each transport lane being opposite. With this arrangement, if any transport lane experiences a fault, such as a collapse or blockage, there is always a transport lane with the same travel direction as the affected lane that can be used as an evacuation lane for the automated driving vehicles. Of course, when many automated driving vehicles are blocked, evacuation lanes with the opposite travel direction to the blocked transport lane can also be used for simultaneous vehicle evacuation. Automated driving vehicles can change direction in the evacuation passage connecting two transport lanes, further increasing evacuation efficiency. Under normal conditions, the automated driving vehicles in the transport lanes travel at a constant speed, and the distance between adjacent automated driving vehicles is a safe distance. During the evacuation process, the autonomous vehicles traveling in the transport lanes, which serve as evacuation routes, are already moving at a safe distance and a preset speed. In order to both continue the original transport tasks of the autonomous vehicles in the evacuation routes to a certain extent and accommodate the autonomous vehicles in the transport lanes where the fault occurred, the autonomous vehicles in the evacuation routes adjust their distance to twice the safe distance when the fault is detected and the evacuation task is issued. When the autonomous vehicles in the evacuation route drive out, they are exactly at the preset speed and inserted between the two autonomous vehicles that have completed the distance adjustment, ensuring the safe driving of the autonomous vehicles.

[0036] In some embodiments of the present invention, the autonomous driving vehicle facing the entrance of the evacuation route adjusts its distance to twice the safe distance, which is achieved in the following way:

[0037] S1. Based on the distance and preset speed in the evacuation channel, determine the first acceleration of the autonomous vehicle that is moving in uniform acceleration in the evacuation channel. The autonomous vehicle at the entrance of the evacuation channel starts to move in uniform acceleration with the first acceleration, so that it reaches the end of the evacuation channel at the preset speed after the evacuation time. The preset speed is the driving speed of the autonomous vehicle in the evacuation channel.

[0038] S2, when an autonomous vehicle enters the evacuation passage and begins to accelerate with the first acceleration, prediction begins. After predicting the evacuation time, the autonomous vehicle that has not passed the preset location and is closest to the preset location in the evacuation passage is marked as the target vehicle; where the preset location is the position in the evacuation passage along the direction of oncoming traffic, at a safe distance from the entrance of the evacuation passage.

[0039] S3, the target vehicle determines the second acceleration and the third acceleration based on its current position and the distance to the preset location. During the evacuation time, the target vehicle adjusts its speed by first decelerating and then accelerating with the second acceleration and the third acceleration in sequence, so that the target vehicle reaches the preset location at the preset speed.

[0040] S4, after the preceding autonomous vehicle enters the evacuation channel, the autonomous vehicle in the transport tunnel where the fault occurred waits at the entrance of the evacuation channel. After the preceding autonomous vehicle exits the evacuation channel, it begins to accelerate with an initial velocity of 0 and a first acceleration, repeating S2-S3.

[0041] In this embodiment, the first acceleration can be determined based on the predetermined evacuation time and distance in the evacuation channel, combined with a preset speed. Specifically, the autonomous vehicle starts from one end (channel entrance) of the blocked transport lane with an initial velocity of 0, accelerates using the evacuation time as the acceleration time, and uses the first acceleration as the acceleration. After uniform acceleration, it reaches the other end (end) of the evacuation channel, at which point its speed is exactly the preset speed, allowing it to synchronize with other autonomous vehicles traveling at the same preset speed in the evacuation lane and perfectly insert itself into the convoy. To prevent the autonomous vehicle that rushes out of the evacuation channel from colliding with the autonomous vehicles that were originally in the evacuation lane, and to maintain a good distance after insertion, the distance between the autonomous vehicles in the evacuation lane needs to be adjusted to twice the safe distance. After adjustment, the autonomous vehicle that exits from the evacuation channel inserts itself between the two vehicles at the preset speed. Specifically, after predicting the evacuation time, the autonomous vehicle closest to the channel entrance but not yet past the entrance is designated as the target vehicle. To maintain transport function and save energy, the autonomous vehicles in front of the target vehicle continue to drive normally without needing to adjust their distance. The target vehicle and the autonomous vehicles behind it need to be spaced out to ensure that each autonomous vehicle exiting the evacuation lane is positioned precisely between two vehicles in the evacuation alley. Segmented control is implemented with the target vehicle as the target, causing it to decelerate and then accelerate within the evacuation time, eventually returning to a preset speed, and reaching the preset location after the evacuation time. Timing begins when the autonomous vehicle starts moving from zero speed at the evacuation lane entrance. The current distance of the target vehicle from the preset location is used as a calculation parameter, combined with the evacuation time and preset speed, to determine the second and third accelerations for deceleration and acceleration, ultimately achieving orderly evacuation. After the preceding vehicle exits the lane, the autonomous vehicles waiting at the evacuation lane entrance begin to accelerate and repeat steps S2-S3 until all autonomous vehicles have completed their evacuation.

[0042] It should be noted that, in order to save energy and allow the autonomous vehicles in the evacuation lanes to continue their transportation tasks, it is not necessary to adjust the distance between each vehicle. Only after the target vehicle is calculated and determined, the distance between the target vehicle and the vehicle in front can be adjusted.

[0043] In this embodiment, the first acceleration a0 can be determined based on the distance S of the evacuation route and the preset speed v0, specifically by the formula v0. 2 =2a0S. Evacuation time T can be calculated using S=0.5(a0T). 2 )Sure.

[0044] In some embodiments of the present invention, the absolute values ​​of the second acceleration and the third acceleration are equal, their directions are opposite, and the durations of the second acceleration and the third acceleration are the same, both being half the evacuation time.

[0045] In some embodiments of the present invention, the absolute values ​​of the second acceleration and the third acceleration are both 4(v0T-S) / T. 2 Where v0 is the preset speed, T is the evacuation time, and S is the distance between the target vehicle and the preset location.

[0046] In this embodiment, it can be verified based on the energy consumption model and its constraints that when the absolute values ​​of the second and third accelerations are both 4(v0T-S) / T 2 At that time, the entire evacuation process is the most energy-efficient.

[0047] The verification process is as follows:

[0048] Energy consumption model:

[0049] The total time formula is T = t1 + t2. The total distance traveled after deceleration and acceleration is S = v0T - (ab × T). 2 ) / 2(a + b);

[0050] The final velocity of the deceleration phase is v1 = v0 - at1, and the final velocity of the acceleration phase is v0 = v1 + bt2. By first decelerating and then accelerating, v0 remains constant, i.e., at1 = bt2. Substituting these values ​​into the total time formula, we get t1 = Tb / (a + b) and t2 = Ta / (a + b).

[0051] During uniformly accelerated motion, the energy consumption E is directly proportional to the sum of the squares of the accelerations and the time, that is:

[0052] E∝a 2 t1+b 2 t2;

[0053] Substituting the time relationships t1 = Tb / (a+b) and t2 = Ta / (a+b), we get:

[0054] E∝a 2 (Tb / (a+b))+b 2 (Ta / (a+b)) = Tab;

[0055] Therefore, minimizing E is equivalent to minimizing ab;

[0056] Where v0 is the preset velocity, a and b are the absolute values ​​of the second and third accelerations, respectively, and t1 and t2 are the durations of the second and third accelerations, respectively.

[0057] Restrictions:

[0058] The expression for S is:

[0059] ;

[0060] We need to minimize ab while satisfying this constraint.

[0061] Introducing variables x = a / b and a = x × b, and substituting them into the constraints, we get:

[0062] S=v0T-(xb×b×T 2 ) / 2(xb+b);

[0063] =v0T-(xb 2 T 2 ) / 2b(x+1)= v0T- (xb× T 2 ) / 2(x+1).

[0064] Solve the equation to find b: b = 2(x + 1)(v0T - S) / (x×T) 2 );

[0065] Objective function ab=xb 2 =x×[2(x+1)(v0T-S) / (xT)] 2 )] 2 .

[0066] The objective function has a symmetric solution x=1 within its domain. This symmetric solution is the minimum point of the objective function. That is, when x is a symmetric solution, ab is minimized.

[0067] Symmetry analysis:

[0068] Let a = b, then:

[0069] .

[0070] Solving for: ;

[0071] At this time, ab = a 2 Furthermore, by comparing the asymmetric cases (such as a≠b), it can be proven that the symmetric solution minimizes ab.

[0072] As for how to control an autonomous vehicle to accelerate uniformly, this can be achieved by establishing a vehicle dynamics model and setting a PID control law. The specific steps are as follows:

[0073] 1. Establish a vehicle dynamics model that balances driving force and resistance;

[0074] Vehicle acceleration a m Driven by F drive With driving resistance F resistance The difference determines:

[0075] ;

[0076] in:

[0077] m: Total mass of the vehicle (including load).

[0078] F resistance =F roll +F aero +F grade :

[0079] Rolling resistance: F roll =m roll ×m×g ;

[0080] Air resistance: F aero =0.5×Cd×ρ×A×v 2 ;

[0081] Slope resistance ：F grade =m×g×sin (θ) (θ is the slope angle);

[0082] ;

[0083] T motor For engine / motor output torque, or motor For transmission system / motor efficiency, r wheel For the tire radius, m roll The rolling resistance coefficient, CD The air drag coefficient, r Let be the air density, and A be the frontal area of ​​the autonomous vehicle. v The vehicle speed.

[0084] Set target constant acceleration a desired The required driving force can be deduced from the model:

[0085] F drive = m×a desired +F resistance

[0086] F needs to be updated in real time resistance (e.g., calculated using vehicle speed v and gradient θ).

[0087] 2. Set the PID control law;

[0088] The PID control process is represented as follows:

[0089] ;

[0090] e=a desired −a actual Feedback from the accelerometer a actual For actual acceleration, T command The output torque of the controller;

[0091] Parameter tuning priority: tune first K p Eliminate steady-state error, then adjust K d Suppress overshoot, and finally through K i Eliminate residual errors.

[0092] Feedforward section: Calculates the feedforward control torque T based on the model. ff :

[0093] ;

[0094] Feedback section: PID correction of actual acceleration deviation.

[0095] 3. Unrestricted compensation and dynamic adjustment;

[0096] Air resistance compensation: Air resistance increases with the square of speed, so a feedforward term needs to be added to the controller. T aero_comp :

[0097] ;

[0098] Slope adaptive, updating in real time via slope sensors or GPS elevation data when the slope changes. F grade ;

[0099] If the vehicle load changes (e.g., when carrying cargo), the mass m needs to be estimated in real time using a Kalman filter or least squares method.

[0100] In some embodiments of the present invention, the method further includes:

[0101] Set a constraint that when any autonomous vehicle is less than the minimum distance (preset) from the vehicle in front, the autonomous vehicle will activate adaptive cruise control to ensure that the autonomous vehicle maintains at least the minimum distance from the vehicle in front.

[0102] In this embodiment, the target vehicle's deceleration followed by acceleration to increase its following distance inevitably leads to a decrease in its following distance. To prevent rear-end collisions, a minimum following distance is set. When the following distance between the target vehicle and the vehicle behind is not less than the minimum following distance, the two vehicles are relatively safe and no adjustment is needed. If the following distance is less than the minimum, the vehicle behind activates adaptive cruise control to maintain the minimum following distance at all times. Similarly, the following distance will also change after adaptive cruise control is activated, and the vehicle behind will also be subject to the above constraints.

[0103] In some embodiments of the present invention, adaptive cruise control is implemented using edge computing. Furthermore, vehicle speed adjustment, distance determination, and position determination can also be achieved through edge computing, thus reducing the computing power of the central control center and increasing its data processing speed.

[0104] In some embodiments of the present invention, the evacuation passage between two transport lanes with opposite directions of travel includes, in sequence, an arc-shaped section and a straight section along the vehicle's defined direction of travel.

[0105] In this embodiment, the autonomous vehicles in the evacuation lane initially travel at a slow speed, posing no risk of rollover or skidding when driving on curved sections with large turns. The angle between the straight sections at the rear and the evacuation lane is small, facilitating the merging of autonomous vehicles and stable driving.

[0106] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An artificial intelligence-based automatic driving path planning method in a well, characterized by, The application relates to a safe distance determination method and a safe distance determination device for autonomous vehicles. The safe distance between autonomous vehicles is determined, and the distance between different autonomous vehicles is the safe distance; A plurality of one-way transport channels for autonomous vehicles are planned; the plurality of transport channels are repeatedly arranged in a two-first-direction one-way and two-second-direction one-way mode, so that a transport channel on one side of each transport channel is in the first direction, and a transport channel on the other side of each transport channel is in the second direction; the first direction is an entry direction, and the second direction is an exit direction; each transport channel is one-way connected to the transport channels on its two sides which are limited to different driving directions through at least one evacuation channel; When a fault point occurs in a transport channel and the autonomous vehicles cannot pass through, all the autonomous vehicles in the transport channel are stopped, the autonomous vehicles with their heads facing the channel opening of the evacuation channel enter the nearest evacuation channel in sequence, and if there is no evacuation channel in the direction in which the autonomous vehicles with their heads facing, the autonomous vehicles with their heads facing the channel opening of the evacuation channel enter the evacuation channel in sequence after all the autonomous vehicles with their heads facing the channel opening of the evacuation channel enter the evacuation channel and then drive to the channel opening of the evacuation channel; wherein the first autonomous vehicle enters the evacuation channel and starts timing, the second autonomous vehicle drives to the channel opening of the evacuation channel, and after the timing reaches a preset time, the second autonomous vehicle enters the evacuation channel; When a fault point occurs in a transport channel and the autonomous vehicles cannot pass through, at least one transport channel on the two sides of the transport channel is a dispersion channel, and the autonomous vehicles in the dispersion channel are adjusted to have a distance of two times the safe distance, so that the autonomous vehicles in the dispersion channel are inserted into the autonomous vehicles in the dispersion channel; The autonomous vehicles in the dispersion channel are adjusted to have a distance of two times the safe distance through the following method: S1, a first acceleration of an autonomous vehicle in uniform acceleration motion in the dispersion channel is determined according to the distance in the dispersion channel and a preset speed, the autonomous vehicle at the channel opening of the dispersion channel starts to move at the first acceleration, and reaches the end of the dispersion channel at a preset speed after a dispersion time, the preset speed is the driving speed of the autonomous vehicle in the dispersion channel; S2, when an autonomous vehicle enters the dispersion channel and starts to accelerate at the first acceleration, prediction is started, and the autonomous vehicle in the dispersion channel which is closest to a preset location and has not passed through the preset location is marked as a target vehicle; wherein the preset location is a position in the dispersion channel which is at a distance of the safe distance from the channel opening of the dispersion channel along the direction of incoming vehicles; S3, the target vehicle determines a second acceleration and a third acceleration based on the current position and the distance from the preset location, and the target vehicle adjusts the speed by decelerating and then accelerating at the second acceleration and the third acceleration in sequence within the dispersion time, so that the target vehicle reaches the preset location at the preset speed. S4, the autonomous vehicle in the transport lane of the failure point enters the evacuation passage after the front autonomous vehicle enters the evacuation passage, and waits in the evacuation passage entrance, and after the front autonomous vehicle drives out of the evacuation passage, starts to accelerate at the initial speed 0 with the first acceleration, and repeats S2-S3; The absolute values of the second acceleration and the third acceleration are both 4(v0T-S) / T 2 ; wherein v0 is the preset speed, T is the evacuation time, and S is the distance between the target vehicle and the preset site.

2. The artificial intelligence-based automatic driving path planning method for a mine according to claim 1, characterized in that, The second acceleration and the third acceleration are opposite in direction, and the duration of the second acceleration and the third acceleration is the same, which is half of the evacuation time.

3. The artificial intelligence-based automatic driving path planning method for a mine according to claim 1, characterized in that, Also includes: A constraint condition is set, which is that when the distance between any autonomous vehicle and the front vehicle is less than the minimum vehicle distance, the autonomous vehicle starts adaptive cruise to keep the distance between the autonomous vehicle and the front vehicle at least the minimum vehicle distance.

4. The artificial intelligence-based automatic driving path planning method for a mine according to claim 3, characterized in that, The adaptive cruise is realized by edge computing.

5. The artificial intelligence-based automatic driving path planning method for a mine according to claim 1, characterized in that, The evacuation passage between the two transport lanes with opposite driving directions comprises an arc-shaped section and a straight line section in sequence along the vehicle driving direction.

Images