RU2783780C1 - Method for adaptive control of movement characteristics of highly automated vehicle depending on level of uncertainty in assessment of surrounding situation - Google Patents

Abstract

FIELD: car industry.

SUBSTANCE: invention relates to methods for control of highly automated vehicles (hereinafter – HAV). The method consists in that, during HAV movement, data from wheel encoders and a steering wheel position sensor is sent to an onboard calculator, while the current HAV state and hypotheses on HAV position are formed, as well as a planned movement path in the form of a desired trajectory on a map and a digital map of obstacles. A safe movement speed is determined based on obtained data. For this purpose, a future trajectory is predicted, the resulting trajectory is applied to each hypothesis on HAV position. Next, each of predicted trajectories is checked for the presence of collisions with assessment of the probability of collision with obstacles. The safe speed is selected as the maximum permissible speed with the probability of collision, not exceeding a given threshold. The resulting value of the safe speed is directed to a speed control unit, by means of which control signals are sent to a traction engine and a brake system.

EFFECT: provision of HAV movement with a safe speed.

1 cl, 1 dwg

Description

The invention relates to the field of navigation of unmanned mobile ground systems, and in particular to methods for automatically controlling the movement of highly automated vehicles (VATS) with the ability to control their speed in order to ensure a guaranteed level of traffic safety of the VATS.

In order for VATS to become part of an intelligent transport system, they must guarantee a certain level of security. To achieve this goal, it is necessary to develop control algorithms for VATS, the implementation of which minimizes the probability of collisions between VATS and static and dynamic obstacles in the process of their movement along a given trajectory.

The following technical solutions are known from the art.

There is a known method for controlling an automatic vehicle (ATS) according to patent RU2649962, according to which a mathematically processed image of the roadway with scanning elements is extracted from the long-term memory of the computer, it is mathematically converted into an image in perspective, a real image of the roadway is obtained, a real image of the roadway is mathematically converted and receive images of scan elements. The images of the roadway with scan elements are compared with the mathematically transformed real image of the roadway, and based on the results, control actions are generated for the automatic telephone exchange. The current mathematically processed images of the roadway with scan elements from long-term memory and real images were obtained from at least two linear photosensitive devices. The sequence of actions for controlling the vehicle based on the analysis includes the preparatory stage and the stages of calculating the current steering angle, the current control of the accelerator pedal, determining the presence of obstacles, building an image of the roadway and analyzing its correspondence to the real one, and developing traffic control control actions.

There is known a method for automatically controlling a vehicle (US2019171204), in which a state detection module of a vehicle in question receives vehicle speed data detected by a vehicle speed sensor, acceleration data detected by an acceleration sensor, and steering angle data detected by steering angle sensor, and determines the driving state of the vehicle in question based on these pieces of data. In accordance with the current state, the onboard computer generates control actions for the optimal passage of the trajectory.

The main disadvantage for the presented technical solutions is the lack of consideration of the uncertainty of the vehicle’s own position, which leads to a decrease in the accuracy of determining the position of the vehicle relative to static and dynamic obstacles around the vehicle, as a result of which the control actions generated to control the movement of the vehicle do not ensure the safe movement of the vehicle along this trajectory .

A known method for planning the trajectories of a vehicle in state space from a starting point to an end point with avoidance of a plurality of static and/or dynamic obstacles according to patent US7447593 consists in building a graph of trajectories connecting the starting point to the end point, taking into account the possible position of the obstacles. At the same time, different rules are used to expand, suspend the expansion, and stop the expansion of nodes of the trajectory graph. The resulting trajectories are evaluated against the given constraints. Based on the graph of trajectories, at least one trajectory is obtained that connects the starting point with the final one, satisfying the given restrictions. The vehicle is controlled to move along the obtained trajectory.

The disadvantage of the presented method lies in the fact that in the process of forming the tree of trajectories, the restrictions on the time of trajectory planning, set by the dynamics of the external environment, are not taken into account.

There is a known method for autonomous control of a vehicle (patent US7613553), which consists in the fact that during the movement of the vehicle along the current trajectory in the cycle of collecting and processing information, information is received from sensors, on the basis of which the model of the current state of the external environment is updated, reflecting the presence in the external environment static and dynamic obstacles. Based on the updated model of the current state of the environment, the model of the predicted state of the environment is updated at the moment the autonomous vehicle reaches the end point of the current trajectory. Based on the updated model of the predicted state of the environment, a set of possible trajectories of further movement from the end point of the current movement trajectory is determined. Each trajectory from the generated set of possible trajectories of further movement is compared with a given digital map. From the generated set of possible trajectories of further movement, a trajectory with the maximum expected utility is selected and implemented when the end point of the current trajectory is reached.

This method of autonomous control of a vehicle is characterized by a large amount of calculations, and therefore, for its implementation, there is a need to use a computer with high performance. In addition, this method is focused on a specific type of vehicle - an underwater vehicle. Thus, the implementation of the presented control method as part of unmanned mobile ground systems cannot be widely used.

The method disclosed in the source US7613553, according to the technical essence, is the closest to the claimed invention and can act as a prototype.

The task to be solved by the present invention is to create a method for adaptive control of the motion characteristics of a highly automated vehicle under conditions of uncertainty in the assessments of the surrounding situation and the own position of the VATS based on determining the safe speed of the VATS as the maximum allowable speed of the VATS, for which the probability of collision of the VATS with obstacles during the forecasting period does not exceed a predetermined threshold value, while the probability of a collision of the VATS in the future is estimated by comparing the predicted future trajectories for the entire space of control signals with a digital obstacle map that stores information about the environmental objects in which the VATS moves.

The technical result of the claimed invention is to ensure the movement of the VATS at a safe speed.

As part of the disclosure of the technical essence of the proposed invention, the safe speed of the VATS is understood as the maximum allowable speed of the VATS on the current section of the trajectory, at which the probability of a collision of the VATS in the future with static or dynamic obstacles is negligibly small.

The claimed technical result is achieved by implementing a method for adaptive control of the motion characteristics of a highly automated vehicle, depending on the level of uncertainty in the assessment of the surrounding situation, according to which, during the movement of the VATS, data from the wheel encoders and the steering position sensor are received on the onboard computer of the VATS, based on these data, a the current state of the VATS, at the same time, data from cameras, wheel encoders and other sources of data on the own position of the VATS are received on the on-board computer VATS, based on these data, many hypotheses about the position of the VATS are formed, the planned path of the VATS movement is formed, presented as a sequence points on the map that form the desired trajectory of movement, a digital map of obstacles is formed, then, based on the data received from the presented sources of information of the VATS, the safe speed of the VATS is determined in a cyclic manner, for which the forecast the future trajectory of the VATS movement is checked, its kinematic stability is checked by determining the deviation from the desired trajectory, the obtained trajectory is superimposed on each hypothesis about the position of the VATS, thereby forming a set of hypotheses about the future trajectories of movement, then each of the predicted trajectories is checked for collisions by comparing them with a digital map of obstacles, the probability of collision with obstacles is estimated based on data on future movement trajectories for the current speed value, the safe speed of the VATS is chosen as the maximum allowable speed of the VATS, for which the probability of collision of the VATS during the forecasting period does not exceed a given threshold value, the obtained value of the safe speed of movement VATS is sent to the VATS movement speed control unit, by means of which control signals are set to the traction motor and the brake system, ensuring the VATS movement at a safe speed.

Thus, the implementation of the specified sequence of actions characterizing the claimed method, using the traffic control system of the VATS, allows, under conditions of uncertainty in assessing the surrounding situation in the presence of static and dynamic obstacles, as well as under conditions of uncertainty of the own position of the VATS, in real time to determine with high accuracy of the safe speed of the VATS in this section of the trajectory for the formation of control signals to the traction motor and the brake system, ensuring the safe movement of the VATS in complex and ambiguous traffic situations.

Figure 1 shows a diagram of the system for safe traffic control of the VATS, which implements the proposed method, where 1 is the camera of the VATS, 2 is the wheel encoders, 3 is the rudder position sensor, 4 is the obstacle detector, 5 is the onboard computer, 6 is the digital obstacle map, 7 ‒ module for generating hypotheses about the position of the VATS, 8 ‒ path planning module, 9 ‒ module for adaptive control of the characteristics of the movement of the VATS, 10 ‒ module for predicting the future trajectory, 11 ‒ trajectory control module, 12 ‒ digital model of the VATS, 13 ‒ engine control unit, 14 ‒ traction engine.

The claimed method for adaptive control of the movement characteristics of the VATS is carried out as follows (figure 1).

In the process of moving a highly automated vehicle (HATS) along a given trajectory, the adaptive control module for motion characteristics (9) as part of the on-board computer HATS (5) receives data from wheel encoders (2) and a steering position sensor (3), based on these data, a the current state of the VATS - the current value of the speed of the VATS, the orientation of the VATS. As an additional source of information about the speed of the VATS, a pressure sensor in the brake circuit can act.

At the same time, the VATS position hypotheses module (7) receives data from cameras (1), wheel encoders (2) and other sources of data on the proper position of the VATS (for example, inertial navigation system, GNSS), based on the specified data form a set of hypotheses about the position of the VATS, each of which contains probabilistic information about the position of the VATS in space, including data on the coordinates of the VATS in the coordinate system associated with the map, and the orientation angles of the VATS in space.

To estimate the current position of the VATS, in the preferred embodiment of the proposed method, the Monte Carlo method (particle filter) is used, which allows generating the specified set of hypotheses about the position of the VATS at the current time and the corresponding estimates of the probabilities of the hypotheses.

At the next stage, information about the uncertainty of its own position is sent to the module (9).

By means of the path planning module (8), the planned path of the VATS is formed, presented as a sequence of points on the map, forming the desired trajectory, after which the received data is sent to the module (9).

To take into account the uncertainty in the assessment of the environment, a digital map of obstacles (6) is formed, which can be represented as a previously known map formed before the implementation of the movement of the VATS, with the unchanged boundaries of the carriageway marked on it.

In addition, during the movement of the VATS, images of the environment from the optical cameras (1) are sent to the obstacle detector (4), through which the received images are processed by methods known from the prior art with the identification of moving or stationary objects of the environment around the VATS, that is, obstacles. . Combining information on selected environmental objects, a digital obstacle map (6) can also be formed - a raster or vector representation of obstacles surrounding the VATS.

Further, using the module for adaptive control of the characteristics of the movement of the VATS (9), based on the data obtained from the presented information sources of the VATS (1-3) and modules (6-8), the safe speed of the VATS is cyclically determined.

To do this, the future trajectory of the VATS movement is estimated using the future trajectory prediction module (10), designed to calculate the advanced section of the trajectory in real time based on the dynamic model of the VATS and combines the trajectory control module (11), consisting of a trajectory controller (controllers direction and position deviations), and a speed controller, and a digital dynamic model of the VATS (12). The module (10) receives data on the state of the VATS (the current value of the speed of the VATS, the position and orientation of the VATS), the given path, the speed limit, and the duration of the prediction. The module (12) transmits its current state to the module (11), which, based on these data, generates control signals and sends them to the digital model (12), which, in accordance with the received control signal, changes its state and sends it via feedback to the module ( eleven). These actions are carried out until the sequence of states of the digital model (12) is formed in the module (10), which forms the predicted movement trajectory with a fixed setting of the future movement speed.

Then, the obtained predicted motion trajectory is checked for kinematic stability by determining the deviation from the desired trajectory and sent to the module for adaptive control of the motion characteristics of the VATS (9).

The resulting predicted trajectory is superimposed on each hypothesis about the position of the VATS, forming a set of hypotheses about future trajectories of movement, then each of the predicted trajectories is checked for the presence of collisions with obstacles by comparing these trajectories with a digital map of obstacles (6), based on the results of this comparison, the probabilities of collision with obstacles are determined. static and dynamic obstacles. The future trajectory is considered safe if none of the hypotheses about the position of the VATS on the trajectory intersects with an obstacle.

It is important to note that when determining the probability of collision with static objects, the future trajectory prediction module (10) generates a future trajectory for a probabilistic estimate of the position of the VATS at time t. Then, the profile of the obtained trajectory is subjected to a parallel translation so that its origin would coincide with the current position of each hypothesis about the position of the VATS, and rotated by the corresponding yaw angle. Thus, many hypotheses about future trajectories of movement are formed. The probability of collision with dynamic obstacles is determined in a similar way, except that for each position of the i-th hypothesis at the time t, the future trajectory prediction module (10) forms the future trajectory separately, after which it is the relative profile of each trajectory that is checked for collisions with obstacles.

At the next step, the probability of collision with obstacles is estimated based on data on future motion trajectories for the current speed value, taking into account the probabilities of collisions with static and dynamic obstacles:

- вероятность столкновения с статическими препятствиями;where

- probability of collision with static obstacles;

- вероятность столкновения с динамическими препятствиями.

- probability of collision with dynamic obstacles.

In accordance with the obtained probability of collision with obstacles, the safe speed of the VATS is chosen as the maximum allowable speed of the VATS, for which the probability of collision of the VATS during the forecasting period does not exceed a given threshold value. The maximum allowable speed is determined by the physical properties of the VATS or the terrain along which the route is laid.

To implement the safe adaptive control of the VATS, the obtained value of the safe speed of the VATS is sent to the speed control unit of the VATS, through which control signals are set to the traction motor and the brake system, ensuring the movement of the VATS at a safe speed.

Claims (1)

A method for adaptive control of the movement characteristics of a highly automated vehicle depending on the level of uncertainty in the assessment of the surrounding situation, characterized in that during the movement of a highly automated vehicle (HATS), data from wheel encoders and a steering position sensor are received on the on-board computer of the HATS, based on the specified data form the current state of the VATS, at the same time, data from cameras, wheel encoders and other sources of data on the own position of the VATS are received on the onboard computer of the VATS, based on these data, many hypotheses about the position of the VATS are formed, the planned path of the VATS movement is formed, presented in the form of a sequence points on the map that form the desired trajectory of movement, a digital map of obstacles is formed, then, based on the data received from the presented sources of information of the VATS, the safe speed of the VATS is determined in a cyclic manner, for which it is predicted comfort of the future trajectory of the VATS, check its kinematic stability by determining the deviation from the desired trajectory, impose the resulting trajectory on each hypothesis about the position of the VATS, thereby forming a set of hypotheses about the future trajectories of the movement, then check each of the predicted trajectories for collisions by comparing them with a digital map of obstacles, the probability of collision with obstacles is estimated based on data on future movement trajectories for the current speed value, the safe speed of the VATS is chosen as the maximum allowable speed of the VATS, for which the probability of collision of the VATS during the forecasting period does not exceed a given threshold value, the obtained value of the safe speed of movement VATS is sent to the VATS movement speed control unit, by means of which control signals are set to the traction motor and the brake system, ensuring the VATS movement at a safe speed.

Images