Vehicle risk avoiding method, device, controller, vehicle and product

Abstract

The invention relates to a vehicle risk avoiding method and device, a controller, a vehicle and a product. The method identifies environmental risk information based on an environmental pattern. The method also generates a state tree of the vehicle based on the environmental risk information in response to identifying the environmental risk information. In addition, the method generates a risk avoiding track based on the state tree. The method can be applied to avoiding risks in various road environments and is not limited to a specific road environment, so that the safety of a driver and passengers in a dangerous environment is greatly improved.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the field of intelligent driving, and more particularly, to a method, an apparatus, a controller, a vehicle and a product for automatic evasive trajectory generation for a vehicle. BACKGROUND

[0002] At present, the application of intelligent driving of vehicles is more and more extensive. Intelligent driving is also known as autonomous driving or adaptive driving, which is a technology that uses various sensors (such as cameras, radars, lidar, etc.), advanced computer systems and artificial intelligence algorithms to achieve autonomous operation of vehicles. There are various intelligent driving systems at present. Among various intelligent driving systems, the advanced driver assistance system (ADAS) is becoming more and more popular, and has gradually become the preferred configuration of medium and high-end new energy electric vehicles, and even the standard configuration of some vehicle models. ADAS is an active safety technology that uses various sensors installed on the vehicle to collect environmental data and driving information inside and outside the vehicle at any time, and performs identification, detection and tracking of static and dynamic objects, so as to enable the driver to perceive possible dangers in the shortest time and timely attract the attention of the driver and improve safety.

[0003] At present, the sensors used by ADAS mainly include cameras, lidar, millimeter wave radar and / or ultrasonic waves, etc. These sensors can simultaneously detect light, heat, pressure or other variables for monitoring the state of the vehicle, thereby providing environmental safety information for vehicle safety control. The early ADAS technology mainly focuses on passive warning, which will issue a warning to remind the driver to pay attention to abnormal vehicles or road conditions when the vehicle detects potential dangers. In addition, the sensor group used by ADAS can perceive potential dangers in front of the vehicle, behind the vehicle and around the vehicle, not just in front of the vehicle. For the latest ADAS technology, active intervention to avoid dangers that the driver cannot react to at all becomes more and more common. Therefore, it is possible to apply the powerful perception and analysis capabilities of ADAS to the surrounding environment. SUMMARY

[0004] In a first aspect of embodiments of the present disclosure, a method for vehicle evasive maneuver is provided. The method comprises identifying environmental risk information based on an environmental pattern. The method further comprises generating a state tree of the vehicle based on the environmental risk information in response to identifying the environmental risk information. In addition, the method further generates an evasive trajectory based on the state tree.

[0005] In a second aspect of the embodiments of this disclosure, an apparatus for vehicle hazard avoidance is provided. The apparatus includes an identification unit configured to identify environmental risk information based on an environmental pattern. The apparatus also includes a generation unit configured to generate a state tree of the vehicle based on the identified environmental risk information. Furthermore, the apparatus further includes a hazard avoidance unit configured to generate an hazard avoidance trajectory based on the state tree.

[0006] In a third aspect of embodiments of this disclosure, a controller is provided. The controller includes at least one processor. The controller also includes a memory coupled to the at least one processor and having instructions stored thereon, which, when executed by the at least one processor, cause the controller to perform a method for vehicle hazard avoidance provided according to this disclosure. The method includes recognizing environmental risk information based on environmental patterns. The method further includes generating a state tree of the vehicle based on the environmental risk information in response to recognizing the environmental risk information. Furthermore, the method generates an hazard avoidance trajectory based on the state tree.

[0007] In a fourth aspect of the embodiments of this disclosure, a vehicle is provided. The vehicle includes a vehicle avoidance device as provided in the second aspect of the invention or a controller as provided in the third aspect of the invention.

[0008] In a fifth aspect of embodiments of this disclosure, a computer program product is provided. The computer program product is tangibly stored on a non-transitory computer-readable medium and includes machine-executable instructions that, when executed, cause a machine to implement a method for vehicle hazard avoidance provided according to this disclosure. The method includes identifying environmental risk information based on environmental patterns. The method further includes generating a state tree of the vehicle based on the environmental risk information in response to identifying the environmental risk information. Furthermore, the method generates an hazard avoidance trajectory based on the state tree.

[0009] In a sixth aspect of embodiments of this disclosure, a computer-readable storage medium is provided, wherein machine-executable instructions are stored on the computer-readable medium, which, when executed, cause a machine to implement a method for vehicle hazard avoidance provided according to this disclosure. The method includes identifying environmental risk information based on environmental patterns. The method further includes generating a state tree of the vehicle based on the environmental risk information in response to identifying the environmental risk information. Furthermore, the method also generates an hazard avoidance trajectory based on the state tree.

[0010] It should be understood that the description in the Summary of the Invention section is not intended to limit the key or essential features of the embodiments of this disclosure, nor is it intended to restrict the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0011] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. In the drawings, the same or similar reference numerals denote the same or similar elements, wherein:

[0012] Figure 1 A schematic diagram is shown illustrating the hazardous environments that the vehicle may face according to this disclosure;

[0013] Figure 2 A flowchart of a method for vehicle hazard avoidance according to some embodiments of the present disclosure is shown;

[0014] Figure 3 A schematic diagram of a partial state tree of a vehicle according to some embodiments of the present disclosure is shown;

[0015] Figure 4 A flowchart of a method for vehicle hazard avoidance according to some embodiments of the present disclosure is shown;

[0016] Figure 5 A flowchart of a method for vehicle hazard avoidance according to some other embodiments of the present disclosure is shown;

[0017] Figure 6 A flowchart of a method for vehicle hazard avoidance according to some other embodiments of the present disclosure is shown;

[0018] Figure 7 A block diagram of a vehicle avoidance device according to some embodiments of the present disclosure is shown;

[0019] Figure 8 A flowchart for vehicle avoidance according to some other embodiments of the present disclosure is shown;

[0020] Figure 9 A flowchart for vehicle avoidance according to some other embodiments of the present disclosure is shown;

[0021] Figure 10 A block diagram of an electronic device that can implement several embodiments of the present disclosure is shown. Detailed Implementation

[0022] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure. The embodiments of this disclosure described below with reference to the accompanying drawings are for illustrative purposes only.

[0023] As mentioned above, current vehicle safety avoidance technologies are primarily solutions designed for scenarios with existing road topology. However, with the increasing prevalence of assisted driving in urban environments, existing solutions cannot support situations on open roads and / or without road topology. Furthermore, the avoidance trajectories generated by current solutions are not flexible enough to handle complex scenarios such as urban roads (e.g., urban intersections). Moreover, maps with road topology require high precision, which is very costly to implement. Therefore, a new approach is needed that leverages the perception and analysis capabilities of currently developed intelligent driving systems to mitigate potential risks to drivers.

[0024] Therefore, embodiments of this disclosure propose a method for vehicle hazard avoidance. This method is based on environmental pattern recognition of environmental risk information, and in response to the recognition of the environmental risk information, generates a state tree for the vehicle based on the environmental risk information, and then generates an hazard avoidance trajectory based on the state tree. Finally, the vehicle can use the generated hazard avoidance trajectory to instruct its actuators to execute the hazard avoidance trajectory to perform hazard avoidance.

[0025] In this way, the ability of risk avoidance methods to be applied to real dangerous environments and different road conditions can be improved, thereby significantly reducing the various risks faced by drivers, rather than just being applicable to situations with a clear road topology.

[0026] Figure 1 A schematic diagram illustrating a hazardous environment that the vehicle disclosed herein may face is shown. Figure 1As shown, vehicle 110 is in its own lane, and there is an obstacle 120 in front of it. This obstacle could be a disabled vehicle, such as an overturned or broken-down vehicle, or a large vehicle moving relatively slowly, such as a large truck, or a vehicle with two or more sections. There is also another vehicle 130 behind vehicle 110, approaching it at a relatively high speed. In the overtaking lane to the left of vehicle 110, there is another vehicle 140 preparing to speed past. Therefore, vehicle 110 is in a relatively dangerous environment and requires a method to avoid it. This disclosure provides a method, apparatus, controller, vehicle, and program product that can generate an avoidance trajectory to help vehicle 110 escape this dangerous environment.

[0027] Figure 2 A flowchart of a method 200 for vehicle avoidance according to some embodiments of the present disclosure is shown. The various steps of method 200 can be performed by a controller, device, vehicle, or program product implementing the method. At block 210, environmental risk information is identified based on environmental patterns. These include other vehicles located in the vehicle's lane or on open roads, and the environmental patterns include, but are not limited to: a vehicle approaching from behind 130; a vehicle crossing in the wrong direction; a vehicle that has overturned out of control; or a vehicle suddenly appearing from the front.

[0028] Optionally, for the case of a rear-approaching vehicle 130, the longitudinal speed vs and maximum deceleration of the rear-approaching vehicle 130 can be obtained through the perception and analysis capabilities provided by ADAS. Suppose that the rear-approaching vehicle 130 decelerates with maximum braking capacity, and solve for the minimum longitudinal extrapolation distance S_min (which can be improved by considering the execution system delay and reaction time of other vehicles). If S_min is less than the functionally set safe distance d_min, then the rear-approaching vehicle 130 is identified as posing an environmental risk.

[0029] Optionally, for vehicles using the wrong-way lane, the perception and analysis capabilities provided by ADAS can be used to obtain information such as the lateral speed vl and the maximum deceleration of the vehicle using the wrong-way lane. This allows for the identification of the vehicle's intention to cross or approach oncoming traffic. Based on the vehicle's maximum deceleration capacity in the crossing direction or when approaching oncoming traffic, and its intended trajectory, the trajectory can be extrapolated to determine whether other vehicles can avoid a collision with vehicle 110 before reaching a near-limit stop. If a collision is possible, the vehicle using the wrong-way lane is identified as posing an environmental risk; or...

[0030] Optionally, in the case of an out-of-control overturned vehicle, the ADAS can also be used to obtain the location of the out-of-control overturned vehicle and risk information such as whether it is stationary or still has a certain operational capability, so as to determine whether the out-of-control overturned vehicle has made it impossible for vehicle 110 to drive normally in its own lane. If it has made it impossible for vehicle 110 to drive normally in its own lane, it is identified as an environmental risk of the out-of-control overturned vehicle.

[0031] Optionally, in the event of a vehicle or other traffic participant (such as a pedestrian) suddenly appearing in front, the perception and analysis capabilities provided by ADAS can be used to analyze whether the vehicle can take emergency braking measures to ensure that a collision is avoided without taking lateral actions, based on the vehicle's state, system delay, and maximum braking capacity. If the vehicle cannot ensure that a collision occurs without taking lateral actions, then the traffic participant that suddenly appears in front is identified as a potential risk.

[0032] It should be noted that environmental risk information based on environmental pattern recognition can be achieved in many ways, such as through pattern recognition using artificial intelligence (AI) or by identifying scenarios according to predetermined rules. The specific method of identifying environmental risks does not constitute any limitation on the scope of protection of this disclosure. Those skilled in the art can use any applicable technology to implement the identification steps of this disclosure.

[0033] At box 220, in response to the identification of the environmental risk information, a state tree of the vehicle is generated based on the environmental risk information. The vehicle state tree can refer to a combination of possible choices made by the vehicle in a certain state regarding hazard avoidance, such as directional selection, speed / acceleration, and time. Figure 3 A schematic diagram of a partial state tree 300 of a vehicle according to some embodiments of the present disclosure is shown, which will be further described below.

[0034] In box 230, an avoidance trajectory is generated based on the state tree. The state tree provides a selection of avoidance strategies, and an avoidance trajectory can be generated based on the selected avoidance strategy. Based on the generated avoidance trajectory, the vehicle 110 can send the avoidance trajectory to the vehicle's actuators to perform an emergency avoidance action.

[0035] Therefore, the above-described embodiments, by comprehensively considering real road conditions and the associated risks, can provide drivers with relatively effective risk avoidance methods, thereby providing drivers with a relatively sufficient guarantee of safe driving. Furthermore, the method 200 of this disclosure embodiment can be applied to roads with complex conditions, and is not limited to roads with a defined road topology.

[0036] Figure 4A flowchart of a method 400 for vehicle hazard avoidance according to some embodiments of the present disclosure is shown. Method 400 begins at block 405. At block 410, it is detected whether the vehicle 110 is stopped due to a malfunction of a vehicle ahead. If yes, the check ends and proceeds to block 440; if no, proceeding to block 420, it is detected whether the vehicle 110 is following a large vehicle ahead at low speed at close range. If yes, the check ends and proceeds to block 440; if no, proceeding to block 430, it is detected whether there is a vehicle blocking the lane where the vehicle 110 is located, making a comfortable lane change unsuitable. If yes, the check ends and proceeds to block 440. If no, it indicates that the vehicle 110 is in a normal driving state, and proceeds to block 450. At block 440, the vehicle needs to enter a vehicle hazard avoidance state; in one embodiment, method 200 may be activated. At box 450, if vehicle 110 continues to drive normally and does not need to enter the vehicle emergency avoidance state, it can proceed to box 405 to begin cyclically monitoring the status of vehicle 110, ensuring it can respond to environmental risks and promptly enter the emergency avoidance state. The method 400 is used for vehicle 110 safety status detection. It can cyclically detect, thereby ensuring the safe driving of vehicle 110. When potential risk situations arise, it can promptly utilize the vehicle's existing ADAS perception and analysis capabilities, combined with the method of this disclosure, to select an avoidance trajectory, thus achieving a safer driving experience.

[0037] Figure 5A flowchart of a method for vehicle hazard avoidance according to some embodiments of the present disclosure is shown. The present disclosure uses method 500 in conjunction with some embodiments to illustrate how to generate a state tree for vehicle hazard avoidance. At block 510, generating the state tree of the vehicle includes: generating a lateral strategy for the vehicle based on environmental risk information, wherein the lateral strategy refers to the lateral behavior that the vehicle can choose to avoid risk, such as veering left or right, or maintaining the current lateral movement of the vehicle, i.e., not changing its original lane. Based on the environmental risk of the vehicle, for structured road scenarios with road topology, the possibility of the vehicle maintaining its current lane or turning to the left or right lane to complete the hazard avoidance action can be determined. For open scenarios (e.g., urban road intersections, suburban roads, etc.), the lateral possibility of the vehicle is determined according to a pre-defined lateral movement space required for the vehicle to complete the hazard avoidance. These lateral strategies for hazard avoidance can be denoted as L_solution: {L_center, L_left, L_right}, where L_solution represents the vehicle's alternative lateral strategies for hazard avoidance; L_center indicates that the vehicle remains in its original lane and does not change its direction of travel; L_left indicates that the vehicle needs to veer left into the left lane (if there is a left lane, or an open area on the left) to avoid environmental risks; and L_right indicates that the vehicle needs to veer right into the right lane (if there is a right lane, or an open area on the right) to avoid environmental risks. Of course, the three behaviors of L_solution mentioned above are semantic-level. Those skilled in the art can make more detailed classifications of the orientation of L_solution based on practical engineering experience or knowledge obtained through artificial intelligence (AI) simulations. These more detailed classifications are also within the scope of protection of this disclosure. Alternatively, the validity of the generated L_solution can be checked using the intent information of potential risk obstacles obtained by ADAS, and invalid lateral strategies can be eliminated. This can significantly improve computational efficiency for subsequent calculations, increase the application's response speed, and save a significant amount of computational costs. For example, by analyzing the intent information of potential risk obstacles, it is known that the risk obstacle in the lane is also turning to the left lane at a high speed while the right lane is safe and free of risk obstacles. In this case, the possible behavior L_left in the lateral strategy can be pruned in advance. This is because if the strategy of turning to the left lane is still maintained at this time, it will put the vehicle that is trying to avoid danger at great risk. The evaluation cost of taking the left-side avoidance behavior is obviously higher than the cost of the right-side avoidance behavior. Early pruning improves computational efficiency and saves computational costs.

[0038] At box 520, generating the vehicle's state tree further includes generating a combination of longitudinal strategies for the vehicle based on the lateral strategies. For each possible lateral strategy, a search is performed to generate a combination of discrete longitudinal strategies for the vehicle within a future fixed time period Ts, {S_t1, S_t2, S_t3, S_t4…S_ti…S_ts}, where S_ti represents the longitudinal strategy adopted by the vehicle for hazard avoidance during time period ti, i is a natural number, and time period ti can be measured in milliseconds, seconds, or minutes. However, excessively long time periods are not suitable because maintaining a fixed strategy and reaction for an extended period during an emergency hazard avoidance event, without changing over time, can easily lead to risks. The longitudinal strategy refers to the vehicle's possible movement in the direction of travel to avoid risks, such as maintaining the current speed, accelerating, or decelerating. For simplicity, the longitudinal action selection combination generation example is set as S_total within the avoidance function: {S_action0, S_action_1, S_action_2}. Here, S_total represents the longitudinal action options available to the vehicle; S_action0 can be set to maintain the current speed; S_action_1 indicates the vehicle needs to accelerate; and S_action_2 indicates the vehicle needs to decelerate. Of course, those skilled in the art will understand that the above settings are merely exemplary, and other different settings are possible. As long as these settings are applicable to this disclosure, they still fall within the scope of protection of this disclosure. If the vehicle selects the longitudinal action S_action0 and executes it at the initial moment, the vehicle's selectable action range at the next state transition moment is S_total. All longitudinal action combinations under this lateral strategy are generated by sequentially selecting actions discretely within the trajectory generation range. Figure 3 A schematic diagram of a partial state tree of a vehicle according to some embodiments of the present disclosure is shown. It illustrates the longitudinal strategies employed by the vehicle during time periods t0, t1, t2, and t3 when the lateral strategy is L_left (i.e., the vehicle's lateral strategy is left-turning). These possible combinations of longitudinal strategies constitute a part of the vehicle's state tree, which can be referred to as... Figure 3 The state tree of the vehicle shown is a part of the vehicle's sub-state tree.

[0039] At box 530, generating the vehicle's state tree further includes: generating alternative avoidance trajectories based on the combination of the lateral and longitudinal strategies. Combining the vehicle's lateral and longitudinal strategies generates multiple alternative trajectories, such as... Figure 3The vehicle's sub-state tree is shown, and finally, a unified state tree is formed. This unified state tree represents all the alternative avoidance behaviors, which can be used to generate the vehicle's subsequent avoidance trajectory. A combination of, but not limited to, lateral and longitudinal movements can be used. For discrete longitudinal strategies, a desired speed or desired acceleration can be set to match the vehicle's risk environment. For lateral strategies, the vehicle reaches the corresponding lane or undergoes lateral displacement when road topology exists. The lateral and longitudinal strategies can be converted into the actual lateral and longitudinal movement information required by the planning algorithm. Furthermore, it is necessary to analyze the potential risks arising from the adopted lateral and longitudinal movements. By comprehensively considering these factors, the final avoidance trajectory that the vehicle can execute can be generated. The relationship between these factors can be represented by the following formula:

[0040] Lateral_information = Fl(L_solution) (1)

[0041] Among them, the lateral strategy L_solution is mapped to the actual lateral information Lateral_information required by the planning algorithm through the function Fl;

[0042] Longitudinal_information = Fs(S_action) (2)

[0043] Among them, based on the vertical strategy S_action, the function Fs maps to the actual vertical information Longitudinal_information required by the algorithm.

[0044] Avoidance_trajectory=Fc(Lateral_information,longitudinal_information,Potential_risk_information)(3)

[0045] The Lateral_information, Longitudinal_information, and Potential_risk_information obtained from equations (1) and (2), along with the resulting potential risk information, are used to generate an executable avoidance trajectory for the vehicle via the function Fc. Generally, due to the need to consider the urgency of the scenario, trajectory generation requires the ability to directly generate an executable avoidance trajectory for the vehicle via the function Fc. The generated trajectory result is then transmitted to the vehicle's actuators. Therefore, the generation algorithm needs to consider the actual vehicle response limit state, and the effectiveness of the generated vehicle avoidance trajectory needs to be checked for its avoidance purpose. Functions Fl, Fs, and Fc can be implemented using existing planning algorithms in the industry, or they can be directly generated by training an AI model. Any method suitable for this disclosure is acceptable, and this disclosure does not impose any restrictions on them. Any form of implementation function used falls within the protection scope of this disclosure.

[0046] It is worth noting that boxes 510, 520, and 530 do not need to be implemented in the same module; each can be an independent step or module that runs autonomously. Figure 5 The terms "integrated" and "together" are used for ease of explanation only, hence the use of boxes to distinguish them. The advantage of the method 500 described in this disclosure is that it can provide hazard avoidance solutions for complex roads, not just those with a defined road topology. Furthermore, method 500 fully considers the environmental risks of the vehicle, thus taking into account various possible hazard avoidance strategies and providing more safe options, preventing the loss of good hazard avoidance opportunities due to adhering to a single strategy. In addition, this method also comprehensively considers various environmental risk factors in the surrounding environment, thereby providing more comprehensive safety protection for the vehicle.

[0047] Figure 6 A flowchart of a method for vehicle hazard avoidance according to some other embodiments of the present disclosure is shown. Figure 6 The method shown, at box 610, further includes generating an avoidance trajectory based on the state tree by performing a risk assessment on the alternative avoidance trajectories. See also Figure 1Taking the risk scenario of vehicle 130 approaching from behind on a structured road as an example, the vehicle is affected by the disabled vehicle 120 (obstacle) in front. After the validity check of the generated trajectory, the vehicle's current lateral strategy is L_left or L_right. For vehicle 110, the behavior of changing lanes to the left is adopted, and traffic participating vehicle 140 is selected as the object of consideration for the search for the optimal solution. The benefit of vehicle 110's risk avoidance behavior to traffic participating vehicle 140 is limited, and in the presence of potential risk scenarios, the probability of traffic participating vehicle 140 sampling the lateral strategy for risk avoidance is low. Therefore, it can be assumed that traffic participating vehicle 140 does not have any lateral strategy change behavior within the risk avoidance action execution interval. For the longitudinal response of traffic participating vehicle 140 to vehicle 110, a simple longitudinal inference model (such as the IDM model shown in equation (4)) can be used to complete the determination of the interaction behavior of traffic participating vehicle 140.

[0048]

[0049] Where in the formula a is the vehicle's acceleration; a and b are the vehicle's maximum acceleration and comfortable deceleration; v0 is the vehicle's free-flow velocity; v is the vehicle's current speed; S0 is the minimum safe following distance; T is the safe following distance; Δv is the speed difference between the vehicle and the vehicle in front; h is the distance between the front ends of the vehicle and the vehicle in front; L is the vehicle length.

[0050] Based on the above behavioral settings, let the reaction trajectory of traffic participant vehicle 140 be obj_traj = F(obj, ego_trajectory), and let the current avoidance trajectory set of vehicle 110 be S_ego. A risk assessment is performed on the possible avoidance trajectories:

[0051] Evaluation_result=WR*Risk_avoidance(ego_trajectory,risk_object)+

[0052] WS*Safety_evaluation(ego_trajectory,traffic_object)+WT*

[0053] Road_rule(ego_trajectory,environment_information) (5)

[0054] in:

[0055] Risk_Avoidance is a risk status assessment of the identified risk obstacles and the vehicle's avoidance trajectory. It can be further designed as a risk field or a specific state mapped to a risk value according to needs. Safety_evaluation is a safety assessment between the vehicle's potential actions and traffic participants. It aims to ensure the safety of traffic participants as much as possible while ensuring the vehicle's own safety avoidance. It involves state quantities and is designed according to needs.

[0056] The Road_rule is used to assess the risk of violations, such as in structured roads where there are solid road lines, and to minimize violations of road restrictions when vehicles need to take evasive action.

[0057] Additionally, WR, WS, and WT are evaluation weight values, which can be formulated by engineers based on historical experience data, or fully consider the driver's risk preferences. Weight values ​​optimized by AI training can also be used, as long as they are applicable to this disclosure, and this does not constitute a limitation of this disclosure. `ego_trajectory` refers to the vehicle avoidance trajectory being evaluated, `risk_object` refers to the identified risk obstacle, `traffic_object` refers to traffic-participating vehicles such as vehicle 140, and `environment_information` refers to the road environment, including road environments with road topology, road intersections, or open road environments. Therefore, optionally, the risk assessment of the candidate avoidance trajectory includes evaluating any one or a combination of the following: obstacle risk assessment between the identified obstacle and the vehicle's avoidance action; safety assessment between the vehicle's avoidance action and other vehicles; or violation risk assessment of road rules. It is worth noting that the evaluation aspects are not limited to the above three aspects. For example, the avoidance efficiency of vehicle 110 can also be considered, i.e., enabling vehicle 110 to pass through the risk area at the fastest speed, which can be used as an evaluation factor for the candidate avoidance trajectory. The above assessment factors are merely illustrative and are not limited to this disclosure. Other risk assessments that are applicable to this disclosure should also be within the scope of protection of this disclosure.

[0058] At box 620, generating the avoidance trajectory based on the state tree further includes: selecting the minimum-risk avoidance trajectory based on the risk assessment results. Specifically, based on the assessment of different avoidance trajectories for the vehicle, the optimal trajectory obtained from the search is selected as the optimal solution for the vehicle's avoidance trajectory in the current scenario. The minimum-risk avoidance trajectory is the minimum-risk avoidance trajectory after a comprehensive assessment of the obstacle risk assessment, safety assessment, and violation risk assessment. Then, the selected minimum-risk avoidance trajectory is sent to the vehicle's actuator for avoidance. Compared to the traditional method of obtaining the optimal vehicle action based on optimal lane change and optimal longitudinal expected vehicle speed, method 600 generates a more flexible longitudinal vehicle action within the trajectory, which is closer to human driver reactions. Furthermore, method 600 can consider that dangerous scenarios can break through road restrictions, such as crossing solid lines, to complete risk avoidance. Additionally, method 600 is not limited to structured roads, thereby reducing the vehicle's dependence on road-level information and improving the applicability of the solution.

[0059] Figure 7 A block diagram of a vehicle avoidance device 700 according to some embodiments of the present disclosure is shown. The device includes an identification unit 710 configured to identify environmental risk information based on environmental patterns. The device 700 also includes a generation unit 720 configured to generate a state tree of the vehicle based on the identified environmental risk information. The device 700 further includes an avoidance unit 730 configured to generate an avoidance trajectory based on the state tree. The advantage of using the method 700 of the present disclosure is that it can provide avoidance solutions for complex roads, not just those with a defined road topology. Furthermore, the method takes into full account various environmental risk factors in the surrounding environment, thereby providing more comprehensive safety protection for the vehicle.

[0060] In some embodiments, the vehicle enters a vehicle avoidance state when the vehicle is in one or a combination of the following situations: the vehicle is stopped due to a malfunction of a vehicle ahead; the vehicle is following a large vehicle ahead; or there is a traffic jam in the lane where the vehicle is located and it is not suitable to change lanes.

[0061] In some embodiments, where other vehicles are located in the vehicle's lane or on an open road, the environmental mode includes, but is not limited to, at least one or a combination thereof: a vehicle approaching from behind; a vehicle traveling in the wrong direction; or a vehicle that has overturned out of control.

[0062] In some embodiments, the generation unit 720 is further configured to generate a lateral strategy for the vehicle based on environmental risk information. Optionally, the generation unit 720 is further configured to generate a combination of longitudinal strategies for the vehicle based on the lateral strategy. Optionally, the lateral strategy is further validated based on the intent information of potential risk obstacles in the environmental risk information, and invalid lateral strategies are deleted.

[0063] In some embodiments, the risk avoidance unit 730 is further configured to generate alternative risk avoidance trajectories based on the combination of the lateral strategy and the longitudinal strategy. Optionally, the risk avoidance unit 730 is further configured to perform a risk assessment on the alternative risk avoidance trajectories and, based on the result of the risk assessment, select the risk-minimizing risk avoidance trajectory.

[0064] In some embodiments, the risk assessment of the alternative avoidance trajectory includes assessing any one or a combination of the following: an obstacle risk assessment between the identified obstacle and the vehicle's avoidance maneuver; a safety assessment between the vehicle's avoidance maneuver and other vehicles; or a violation risk assessment of road rules. Optionally, the minimum risk avoidance trajectory is a minimum risk avoidance trajectory obtained after a comprehensive assessment of the above obstacle risk assessment, safety assessment, and violation risk assessment. Optionally, the device 700 also sends the selected minimum risk avoidance trajectory to the vehicle's actuator for avoidance.

[0065] This disclosure also provides a controller comprising: at least one processor; and a memory coupled to the at least one processor and having instructions stored thereon, the instructions, when executed by the at least one processor, causing the controller to perform a method for vehicle avoidance according to this disclosure. The method includes recognizing environmental risk information based on environmental patterns. The method further includes generating a state tree of the vehicle based on the environmental risk information in response to recognizing the environmental risk information. Furthermore, the method also generates an avoidance trajectory based on the state tree. Additionally, the instructions, when executed by the at least one processor, cause the controller to perform methods provided according to other embodiments of this disclosure.

[0066] This disclosure also provides a vehicle that includes the above-described device 700 and / or other devices disclosed in related embodiments.

[0067] This disclosure also provides a computer program product tangibly stored on a non-transitory computer-readable medium and comprising machine-executable instructions that, when executed, cause a machine to perform a method for vehicle avoidance according to this disclosure, when executed by the at least one processor. The method includes recognizing environmental risk information based on environmental patterns. The method further includes generating a state tree of the vehicle based on the environmental risk information in response to recognizing the environmental risk information. Furthermore, the method generates an avoidance trajectory based on the state tree. Additionally, when executed by the at least one processor, the instructions cause the machine to perform methods provided according to other embodiments of this disclosure.

[0068] Figure 8 A flowchart for vehicle avoidance according to some other embodiments of the present disclosure is shown. At block 805, method 800 for vehicle avoidance begins. At block 810, it is checked whether the vehicle is in a situation where the avoidance function is potentially active. If not, the process ends at block 850; if so, the process proceeds to block 820. Optionally, method 400 is used to check whether the vehicle is in a situation where the avoidance function is potentially active. At block 820, potential risks are identified. If no potential risk exists, the process ends at block 850; if a potential risk exists, the process proceeds to block 830. At block 830, the vehicle generates an avoidance trajectory based on environmental risk information. If an avoidance trajectory can be generated, optionally, methods 500 and 600 are used to generate the trajectory, and the trajectory is output to block 840; if an avoidance trajectory cannot be generated, the process ends at block 850. At block 840, the vehicle's actuators perform avoidance actions based on the avoidance trajectory, and the process ends at block 850.

[0069] Figure 9 A flowchart for vehicle avoidance according to some other embodiments of the present disclosure is shown. At block 910, method 900 for vehicle avoidance begins. At blocks 920, 930, and 940, optionally, a lateral strategy, a combination of longitudinal strategies, and candidate avoidance trajectories for the vehicle are generated according to method 500, respectively. At block 950, candidate avoidance trajectories are searched; optionally, an optimal avoidance trajectory can be generated according to method 600. At block 960, avoidance actions are performed using the vehicle's actuators based on the optimal avoidance trajectory, and the process ends at block 950.

[0070] Figure 10A block diagram of a controller 1000 that can implement various embodiments of the present disclosure is shown. As shown, the electronic device 1000 includes a processor 1001, which can perform various appropriate actions and processes according to computer program instructions loaded into random access memory (RAM) 1003 based on computer program instructions stored in read-only memory (ROM) 1002. Various programs and data required for the operation of the electronic device 1000 may also be stored in RAM 1003. The processor 1001, ROM 1002, and RAM 1003 are interconnected via bus 1004. An input / output (I / O) interface 1005 is also connected to bus 1004.

[0071] Processor 1001 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 1001 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 1001 performs the various methods and processes described above, such as method 200. For example, in some embodiments, method 200 may be implemented as a computer software program tangibly contained in a machine-readable medium. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 1000 via ROM 1002. When the computer program is loaded into RAM 1003 and executed by processor 1001, one or more steps of method 200 described above may be performed. Alternatively, in other embodiments, processor 1001 may be configured to perform method 200 by any other suitable means (e.g., by means of firmware).

[0072] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload programmable logic devices (CPLDs), and so on.

[0073] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0074] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing. Furthermore, although operations are depicted in a specific order, this should be understood as requiring that such operations be performed in the specific order shown or in sequential order, or requiring that all illustrated operations be performed to achieve the desired result. In certain environments, multitasking and parallel processing may be advantageous. Similarly, while several specific implementation details are included in the foregoing discussion, these should not be construed as limiting the scope of this disclosure. Certain features described in the context of individual embodiments may also be implemented in combination in a single implementation. Conversely, various features described in the context of a single implementation may also be implemented individually or in any suitable sub-combination in multiple implementations.

[0075] Although the subject matter has been described using language specific to structural features and / or methodological logic, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. Rather, the specific features and actions described above are merely illustrative examples of implementing the claims.

Claims

1. A method (200) for vehicle hazard avoidance, comprising: Environmental risk information based on environmental pattern recognition (210); In response to the identification of the environmental risk information, a state tree of the vehicle is generated (220) based on the environmental risk information; as well as Based on the state tree, a (230) avoidance trajectory is generated.

2. The method (200) according to claim 1, wherein the vehicle enters a vehicle evasive state when the vehicle is in at least one of the following conditions: The vehicle was stopped because the vehicle in front of it had broken down; The vehicle was following a large vehicle ahead; or The vehicle is in a lane where there are vehicles blocking the way ahead, making it unsuitable to change lanes.

3. The method (200) of claim 1, wherein other vehicles are located in the vehicle's lane or on an open road, and the environmental mode includes at least one or a combination thereof: Vehicles approaching from behind; Vehicles using the wrong lane; Out-of-control overturned vehicle; or A vehicle suddenly appears in front.

4. The method (200) according to claim 1, wherein generating the state tree of the vehicle comprises: Based on environmental risk information, a lateral strategy for the vehicle is generated.

5. The method (200) according to claim 4, wherein generating the state tree of the vehicle further comprises: Based on the lateral strategy, a combination of longitudinal strategies for the vehicle is generated.

6. The method (200) according to claim 5, wherein generating the avoidance trajectory based on the state tree includes: Based on the combination of the horizontal and vertical strategies, alternative risk avoidance trajectories are generated.

7. The method (200) according to claim 6, wherein generating the avoidance trajectory based on the state tree further includes: A risk assessment is conducted on the proposed alternative evasion trajectories; as well as Based on the results of the risk assessment, the least risky avoidance trajectory is selected.

8. The method (200) of claim 7, wherein risk assessment of the alternative evasion trajectories includes assessing any one or a combination of the following: Obstacle risk assessment between identified obstacles and the vehicle's avoidance maneuvers; A safety assessment of the vehicle's evasive maneuvers relative to other vehicles; or Risk assessment of violations of road rules.

9. The method (200) according to claim 8, wherein the minimum risk avoidance trajectory is the minimum risk avoidance trajectory after comprehensive evaluation of the above obstacle risk assessment, safety assessment and violation risk assessment.

10. The method of claim 4, further comprising: Based on the intent information of potential risk obstacles in the environmental risk information, the effectiveness of the lateral strategies is checked, and invalid lateral strategies are deleted.

11. The method of claim 7, further comprising: The selected least-risk avoidance trajectory is sent to the vehicle's actuators to avoid danger.

12. A device (700) for vehicle avoidance, comprising: The identification unit (710) is configured to identify environmental risk information based on environmental patterns; A generation unit (720) is configured to generate a state tree of the vehicle based on the environmental risk information in response to the identification of the environmental risk information; as well as The avoidance unit (730) is configured to generate an avoidance trajectory based on the state tree.

13. A controller, comprising: At least one processor; as well as A memory coupled to the at least one processor and having instructions stored thereon, which, when executed by the at least one processor, cause the controller to perform the method according to any one of claims 1-11.

14. A vehicle comprising the controller according to claim 13.

15. A computer program product tangibly stored on a non-transitory computer-readable medium and comprising machine-executable instructions that, when executed, cause a machine to perform the method according to any one of claims 1-11.

Images