Automatic driving method, device and electric snowmobile based on radar and vision fusion

Abstract

The embodiment of the present invention discloses an automatic driving method, device and electric snowmobile based on radar-visual fusion, which belongs to the field of automatic driving technology. The embodiment of the present invention can generate multiple initial automatic driving strategies based on the driving user's request information, and mine the task elements that characterize each strategy through the radar-visual fusion state vector, which enables the automatic driving system to comprehensively and meticulously analyze the execution requirements of each strategy under the radar-visual fusion technology. The intelligent driving status assessment record is obtained and the road network environment quality characteristics of each initial strategy are calculated based on this. This process makes full use of past experience data to make the strategy evaluation more reliable and scientific. By screening out the target automatic driving strategy and sending it to the target electric snowmobile, it can ensure that the driving strategy that best suits the current road network environment is selected. The overall adaptability and safety of the electric snowmobile automatic driving are improved.

Description

Technical Field

[0001] The embodiments of the present invention belong to the field of autonomous driving technology, and specifically relate to an autonomous driving method, device, and electric snowmobile based on radar and vision fusion. Background Art

[0002] The field of autonomous driving for electric snowmobiles faces numerous technical challenges. Traditional autonomous driving strategy generation often lacks comprehensive consideration of multiple factors. For example, different driving users have varying needs, but existing technologies may not be able to comprehensively and meticulously generate diverse initial strategies based on these needs. Furthermore, the application of radar-visual fusion technology in autonomous driving lacks an effective method for in-depth analysis of the specific requirements for each strategy execution—that is, the task elements of strategy execution. Furthermore, historical empirical data is not fully utilized when evaluating the suitability of autonomous driving strategies for the current road network environment. This makes it difficult to accurately select the optimal autonomous driving strategy for varying road conditions (such as snow-covered roads, where snow depth, road surface roughness, and width all affect driving), traffic volume (the number of snowmobiles and skiers fluctuates frequently), and weather conditions (inclement weather such as snowfall, low visibility, and strong winds), thereby increasing driving risks. Summary of the Invention

[0003] The embodiments of the present invention provide an automatic driving method, device and electric snowmobile based on radar and vision fusion, which can solve or partially solve the technical problems involved in the above-mentioned background technology.

[0004] An embodiment of the present invention provides an automatic driving method based on radar and vision fusion, which is applied to a radar and vision fusion automatic driving device. The method includes: processing the acquired driving user request information to generate X initial automatic driving strategies for the driving user request information; X is a positive integer; performing radar and vision fusion state vector mining on each of the initial automatic driving strategies to obtain the radar and vision fusion state vector corresponding to each of the initial automatic driving strategies; a radar and vision fusion state vector corresponding to the initial automatic driving strategy is used to characterize the strategy execution task element of the initial automatic driving strategy; obtaining an intelligent driving state evaluation record; the intelligent driving state evaluation record contains a description of the association between a past radar and vision fusion state vector set and a past road network environment quality feature set, and one of the past radar and vision fusion state vector sets There is an association description between a past radar-visual fusion state vector and a past road network environment quality feature in the past road network environment quality feature set, where a past radar-visual fusion state vector refers to a radar-visual fusion state vector corresponding to a past autonomous driving strategy for which a driving user requested information in the past; the road network environment quality feature of each initial autonomous driving strategy is calculated using the association description between the radar-visual fusion state vector corresponding to each initial autonomous driving strategy, the past radar-visual fusion state vector set in the intelligent driving state assessment record, and the past road network environment quality feature set; a target autonomous driving strategy is selected from the X initial autonomous driving strategies based on the road network environment quality features of each initial autonomous driving strategy, and the target autonomous driving strategy is sent to a target electric snowmobile.

[0005] An embodiment of the present invention provides a radar-visual fusion autonomous driving device, comprising at least one processor and a memory; the memory stores computer-executable instructions; the at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor performs the above-mentioned method.

[0006] An embodiment of the present invention provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, the steps of the above method are implemented.

[0007] An embodiment of the present invention provides an electric snowmobile, which is communicatively connected to a radar-vision fusion automatic driving device. The electric snowmobile is used to receive a target automatic driving strategy issued by the radar-vision fusion automatic driving device, and the target automatic driving strategy is determined by the above method.

[0008] In this embodiment of the present invention, multiple initial autonomous driving strategies are generated based on driver-user request information. The task elements representing each strategy are mined using a radar-visual fusion state vector. This enables the autonomous driving system to comprehensively and meticulously analyze the execution requirements of each strategy using radar-visual fusion technology. Intelligent driving state assessment records are obtained and used to calculate the road network quality characteristics for each initial strategy. This process leverages historical empirical data, making strategy evaluation more reliable and scientific. By selecting a target autonomous driving strategy and issuing it to the target electric snowmobile, the most appropriate driving strategy for the current road network environment is selected. This overall improves the adaptability and safety of electric snowmobile autonomous driving. Regardless of road conditions (such as snow depth and road width), traffic flow (vehicle or skier density), and weather conditions (snowfall, visibility, wind speed), the vehicle can make optimal decisions, effectively avoiding driving risks caused by inappropriate strategy selection and improving the user's travel experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 A flowchart of an autonomous driving method based on radar and vision fusion provided by an embodiment of the present invention.

[0010] Figure 2 A schematic diagram of the structure of a radar-visual fusion automatic driving device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0011] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without creative effort shall fall within the scope of protection of the embodiments of the present invention.

[0012] The terms "first," "second," and the like in the embodiments of the present invention are used to distinguish similar objects, and are not used to describe a specific order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments of the present invention can be implemented in an order other than that illustrated or described herein, and that the objects distinguished by "first," "second," and the like are generally of the same type, and do not limit the number of objects; for example, the first object can be one or more. Furthermore, in the embodiments of the present invention, "and / or" refers to at least one of the connected objects, and the character " / " generally indicates that the objects connected are in an "or" relationship.

[0013] Figure 1 An autonomous driving method based on radar and vision fusion is shown, which is applied to a radar and vision fusion autonomous driving device. The method includes the following steps 110 to 150.

[0014] Step 110: The radar-visual fusion autonomous driving device processes the acquired driving user request information and generates X initial autonomous driving strategies based on the driving user request information; X is a positive integer.

[0015] In this embodiment of the present invention, step 110 involves generating an initial autonomous driving strategy. This step focuses on obtaining and processing driver-user request information. The radar-based autonomous driving device must first obtain driver-user request information. This driver-user request information may include various information, such as destination information (e.g., coordinates or a specific location name, such as the coordinates of a specific area at a ski resort (45.321, -73.567)), driving speed requirements (e.g., a maximum speed limit of 50 km / h), and driving mode requirements (e.g., comfort mode, energy-saving mode, etc.).

[0016] The autonomous driving system then processes the received driver request information. This processing may involve operations such as data parsing and format conversion. For example, if the user request information is entered in natural language, such as "I want to go to the rest area of the ski resort in comfort mode," the autonomous driving system needs to convert it into a machine-understandable format and parse out key information such as the destination is the rest area of the ski resort and the driving mode is comfort mode.

[0017] Based on the processed user request information, the radar-assisted autonomous driving device generates X initial autonomous driving strategies. Here, X is a positive integer, and its specific value may depend on various factors, such as different driving routes and speed adjustment strategies. For example, for a request to go to the ski resort rest area, the following three initial autonomous driving strategies may be generated (X = 3).

[0018] Initial autonomous driving strategy 1: Choose the shortest route and drive at the upper limit allowed by traffic regulations, but gradually slow down to 10 km / h when approaching the rest area and maintain a large safety distance (such as 5 meters).

[0019] Initial autonomous driving strategy 2: Choose a route with better scenery but a slightly longer distance, drive at a moderate speed (such as 30 km / h) in energy-saving mode, and set the safety distance to 3 meters.

[0020] Initial autonomous driving strategy three: Choose the route with the least traffic, dynamically adjust the speed based on road conditions (up to a maximum of 40 km / h), and maintain a safe distance of 4 meters.

[0021] Step 120: The radar-visual fusion autonomous driving device mines the radar-visual fusion state vectors of each of the initial autonomous driving strategies to obtain the radar-visual fusion state vectors corresponding to each of the initial autonomous driving strategies; a radar-visual fusion state vector corresponding to an initial autonomous driving strategy is used to represent the strategy execution task elements of the initial autonomous driving strategy.

[0022] Furthermore, step 120 turns to the mining of the radar-visual fusion state vector. The radar-visual fusion state vector is a vector used to characterize the strategy execution task elements of the initial autonomous driving strategy. It contains information on multiple dimensions related to the radar-visual fusion technology, which is of great significance for evaluating the feasibility and effectiveness of the autonomous driving strategy in actual execution. For each initial autonomous driving strategy, the radar-visual fusion autonomous driving device will mine the radar-visual fusion state vector. For example, taking strategy one as an example, the mined radar-visual fusion state vector may contain the following elements:

[0023] LiDAR-related factors: The expected detection range requirement for the LiDAR on the shortest path (e.g., a sector area of 100 meters in front and 10 meters on each side), and the LiDAR resolution requirement (e.g., 0.1 degrees horizontally and 0.5 degrees vertically to ensure accurate detection of small obstacles such as skis).

[0024] Vision sensor-related factors include the target types that the vision sensor needs to identify (such as road signs, other vehicles on snowy terrain, pedestrians, etc.), the required clarity of the visual image (such as a resolution of at least 1080p to accurately identify distant traffic signs), and the sensitivity requirements to different colors and textures (for example, high sensitivity to identifying black objects against a white snow background).

[0025] Fusion-related factors: the time interval for data fusion in the radar-visual fusion algorithm (such as data fusion every 0.1 seconds to ensure timely acquisition of the latest environmental information), the processing priority of the fused data (for example, giving priority to the radar-visual fusion data of objects that are closer to the object to deal with emergencies), etc.

[0026] For other initial autonomous driving strategies (Strategy 2 and Strategy 3), their corresponding radar-visual fusion state vectors will also be mined in a similar manner. These vectors reflect the requirements and dependencies of different strategies on various aspects of radar-visual fusion technology during execution.

[0027] Step 130: The radar-vision fusion autonomous driving device obtains the intelligent driving status assessment record.

[0028] The intelligent driving status assessment record includes a description of the association between a set of past radar-visual fusion state vectors and a set of past road network environment quality features. There is a description of the association between a past radar-visual fusion state vector in the set of past radar-visual fusion state vectors and a past road network environment quality feature in the set of past road network environment quality features. A past radar-visual fusion state vector refers to a radar-visual fusion state vector corresponding to a past autonomous driving strategy for which a past driving user requested information.

[0029] Next, step 130 is used to obtain an intelligent driving state evaluation record. This record contains a description of the association between a set of past radar-visual fusion state vectors and a set of past road network environmental quality characteristics. This record is an important reference for evaluating the newly generated initial autonomous driving strategy.

[0030] Each past radar-visual fusion state vector in the set of past radar-visual fusion state vectors corresponds to a past autonomous driving policy for a driver's request. For example, if a driver requested to travel from the ski resort entrance to the ski lift area, the radar-visual fusion state vector generated for the autonomous driving policy at that time might include factors such as the lidar's detection frequency along that route (one detection every 0.2 seconds) and the required accuracy of the visual sensor for recognizing the lift sign in the snow (at least 90%).

[0031] The past road network environmental quality characteristics in the past road network environmental quality characteristics set reflect the road network conditions at that time. For example, the road snow depth (e.g., average snow depth of 20 cm), road flatness (e.g., with some small undulations and a slope of less than 5%), traffic volume (e.g., 5 other snow-covered vehicles per kilometer), and weather conditions (light snow, visibility of 500 meters).

[0032] The association description establishes a connection between the two. For example, when the lidar detection frequency in the past radar-visual fusion state vector is high and the visual sensor recognition accuracy requirement is high, the corresponding road network environmental quality characteristics may be deep snow and moderate traffic flow, which means that more accurate environmental perception is required to ensure safe driving under such road conditions.

[0033] Step 140: The radar-visual fusion autonomous driving device calculates the road network environment quality characteristics of each of the initial autonomous driving strategies through the radar-visual fusion state vectors corresponding to each of the initial autonomous driving strategies, the association description between the past radar-visual fusion state vector set in the intelligent driving state evaluation record and the past road network environment quality characteristic set.

[0034] Next, step 140 involves calculating the road network environment quality characteristics for the initial autonomous driving strategy. The radar-visual fusion autonomous driving device calculates the road network environment quality characteristics for each initial autonomous driving strategy using the radar-visual fusion state vector corresponding to each initial autonomous driving strategy, the association between the past radar-visual fusion state vector set in the intelligent driving state assessment record, and the past road network environment quality characteristic set.

[0035] For example, if the lidar detection range in the initial autonomous driving strategy 1 is similar to some records in previous intelligent driving state assessments, such as situations with similar detection range requirements, and the corresponding road network quality characteristics in those past situations are shallow snow depth (less than 10 cm), high traffic volume (more than 10 snow-covered vehicles per kilometer), and good road smoothness (slope within 3%), then we can preliminarily infer that strategy 1 may perform well under these road network quality characteristics.

[0036] To achieve more accurate calculations, some algorithmic models can be used. For example, a neural network-based regression model can be used. The initial autonomous driving strategy's radar-visual fusion state vector is used as input, and the association description between the past radar-visual fusion state vectors and the past road network environmental quality feature sets serves as training data. The model's output is the quantified value of the road network environmental quality features of the initial autonomous driving strategy. For example, in a trained neural network model, the number of input layer nodes is determined by the dimension of the radar-visual fusion state vector (e.g., 20 nodes), the hidden layer adopts a three-layer structure (with 30, 20, and 10 nodes per layer, respectively), and the output layer is the quantified value of the road network environmental quality features (e.g., a value between 0 and 1, where 0 indicates extremely poor road network environmental adaptability and 1 indicates excellent road network environmental adaptability).

[0037] The same method is used to calculate the road network environment quality characteristic quantification values for the other initial autonomous driving strategies (Strategy 2 and Strategy 3). For example, the road network environment quality characteristic quantification value calculated for Strategy 2 is 0.6, and the quantification value obtained for Strategy 3 is 0.7.

[0038] Step 150: The radar-visual fusion autonomous driving device selects a target autonomous driving strategy from the X initial autonomous driving strategies based on the road network environment quality characteristics of each initial autonomous driving strategy, and sends the target autonomous driving strategy to the target electric snowmobile.

[0039] Finally, step 150 is used to select and issue the target autonomous driving strategy. The radar-vision fusion autonomous driving device selects the target autonomous driving strategy from the X initial autonomous driving strategies based on the road network quality characteristics of each initial autonomous driving strategy. This selection process is based on the quantitative values of the road network quality characteristics calculated previously.

[0040] For example, when the quantitative values of the road network environment quality characteristics of strategies one, two, and three calculated above are 0.5, 0.6, and 0.7, respectively, since the quantitative value of the road network environment quality characteristics of strategy three is the highest, strategy three is selected as the target autonomous driving strategy.

[0041] Once the target autonomous driving strategy (Strategy 3 in this case) is selected, the Leishi Fusion autonomous driving device will send it to the target electric snowmobile. Upon receiving this strategy, the target electric snowmobile will then operate according to the requirements of Strategy 3, including following the selected route with the least traffic, dynamically adjusting speed based on road conditions (up to a maximum of 40 km / h), and maintaining a safe distance of 4 meters.

[0042] In summary, through a series of steps such as processing the driver's request information by the radar-vision fusion autonomous driving device, mining the radar-vision fusion state vector, utilizing the intelligent driving status assessment records, and calculating the road network environment quality characteristics, the target autonomous driving strategy that best suits the current situation is finally selected and issued to the target electric snowmobile, thereby improving the autonomous driving performance and safety of the electric snowmobile with the support of radar-vision fusion technology.

[0043] In the aforementioned technical solution, the radar-visual fusion state vector is a multi-dimensional vector that integrates various state information of the lidar and vision sensors during the execution of the autonomous driving strategy. Each dimension of this vector represents a specific attribute related to the radar-visual fusion technology.

[0044] From the perspective of lidar, the vector may include dimensions related to the lidar's detection range. For example, in the case of an electric snowmobile, these dimensions may include forward detection distance (e.g., in meters, with values ranging from tens to hundreds of meters, such as 150 meters forward) and side detection width (e.g., 15 meters on each side). It may also include dimensions related to the lidar's resolution, such as horizontal resolution (which can be between 0.1 and 1 degrees, such as 0.3 degrees) and vertical resolution (e.g., 0.5 degrees), which directly impact obstacle detection accuracy. Furthermore, the lidar's detection frequency (e.g., every 0.1 to 1 second, such as every 0.2 seconds) is also a dimension of the vector, which determines the speed at which environmental information is updated.

[0045] For visual sensors, the radar-visual fusion state vector includes dimensions related to visual target recognition. In an electric snowmobile driving scenario, these might include target type (e.g., categorizing targets into traffic signs, other vehicles, pedestrians, and special snow markings, where all types may need to be recognized), sensitivity to different colors and textures (e.g., higher sensitivity to black objects against a white snow background, represented by a value between 0 and 1, with 0.8 indicating higher sensitivity), and more. The clarity requirement for the visual image is also a dimension, expressed in pixel resolution (e.g., at least 720p or even higher, 1080p).

[0046] In terms of radar-visual fusion, the vector includes the time interval dimension of data fusion (such as data fusion every 0.05 seconds to 0.5 seconds, for example, every 0.1 seconds), the processing priority dimension of fused data (for example, targets can be divided into three levels: near, medium, and far according to distance, and the fused data of targets with close distances are processed first, with 1 representing the highest priority), etc.

[0047] The primary function of the radar-visual fusion state vector is to comprehensively and accurately describe the requirements and characteristics of an initial autonomous driving strategy at the radar-visual fusion technical level. It provides a detailed basis for subsequent evaluation of the autonomous driving strategy's adaptability in various road network environments. This vector clearly demonstrates how the lidar and vision sensors must work together and what performance levels each must achieve when executing a specific autonomous driving strategy. For example, in an autonomous driving strategy that requires frequent avoidance of skiers, the values of dimensions such as the lidar detection frequency and the vision sensor's sensitivity to pedestrian (skier) recognition in the radar-visual fusion state vector will be adjusted accordingly to ensure timely and accurate perception of the skier's position and intended movements, thereby ensuring safe operation of the electric snowmobile.

[0048] In addition, the strategy execution task elements are the key components involved in the execution of the initial autonomous driving strategy. These elements can be roughly divided into several categories, such as driving route related elements, speed control related elements, and safety assurance related elements.

[0049] The elements related to the driving route include the determination of the destination (e.g., expressed in coordinate form, the specific coordinates of the ski resort (42.123, -75.456) are the destination), the selection of the driving route (e.g., different selection strategies such as choosing the shortest path, the path with better scenery, or the path with the least traffic).

[0050] Speed control-related factors include the speed setting mode (such as fixed speed mode, set at 30 km / h; or dynamic speed mode, dynamically adjusting the speed according to road conditions and traffic regulations, for example, the maximum speed on an open snowy road does not exceed 50 km / h, and the speed is reduced to 10 km / h when approaching a crowded area), and the basis for speed adjustment (such as based on the distance to the vehicle or pedestrian in front, the curvature of the road, or traffic signs).

[0051] Safety assurance-related factors include the setting of safe distances (such as maintaining a safe distance of 3 meters from the vehicle in front during normal driving, and increasing it to 5 meters under special road or weather conditions), and emergency response strategies (for example, when a skier is suddenly detected sliding quickly across the snowmobile's route, should emergency braking or evasive measures be taken, and how to determine the direction and magnitude of the evasion, etc.).

[0052] Furthermore, the radar-visual fusion state vector maps and supports the strategic execution mission elements from the perspective of radar-visual fusion technology. For example, the route selected in the strategic execution mission element affects the lidar detection range and target type recognition of the visual sensor in the radar-visual fusion state vector. If a route with heavy traffic is selected, the lidar detection range in the radar-visual fusion state vector may need to be wider to detect more vehicles in advance; the visual sensor needs to focus more on vehicle type recognition, including distinguishing between snowmobiles of different sizes and uses. Similarly, speed control-related elements are also related to the radar-visual fusion state vector. At high speeds, the lidar detection frequency in the radar-visual fusion state vector may need to be higher, and the visual sensor image clarity must also be higher to ensure timely and accurate perception of environmental changes at high speeds to ensure safety. The safety distance setting in the safety assurance element is related to the detection accuracy in the radar-visual fusion state vector. More accurate radar-visual fusion detection allows for more reasonable determination of the safety distance.

[0053] In the embodiment of the present invention, the road network environmental quality characteristics are a comprehensive description of the road network environmental conditions on which the electric snowmobile travels, and are mainly composed of the physical characteristics of the road, traffic flow characteristics, and weather condition characteristics.

[0054] The physical characteristics of the road include snow depth (in centimeters, ranging from a few centimeters to tens of centimeters, such as 15 centimeters), road flatness (expressed as a percentage of slope, such as a 2% slope indicates a relatively flat road), road width (for example, a single lane width of 3 meters or a double lane width of 6 meters, etc.), etc.

[0055] Traffic flow characteristics refer to the number of vehicles and pedestrians (including skiers) on the road. For example, the number of snowmobiles per kilometer of road (e.g., 8 snowmobiles per kilometer) or the density of pedestrians (skiers) (e.g., 20 skiers per square kilometer).

[0056] Weather condition characteristics include snowfall (such as light snow, moderate snow or heavy snow, which can be quantified as snowfall per hour, such as light snow is 0.1 mm snowfall per hour), visibility (in meters, such as visibility is 300 meters), wind speed (in meters / second, such as wind speed is 5 meters / second), etc.

[0057] These features can be represented quantitatively. For example, values such as snow depth, traffic flow, and snowfall can be normalized to a range between 0 and 1. For snow depth, 0 indicates no snow, and 1 indicates a preset maximum snow depth (e.g., 50 cm). A traffic flow value of 0 indicates no vehicles or pedestrians, and 1 indicates saturated traffic flow.

[0058] Road network quality characteristics play a key role in assessing the suitability of an initial autonomous driving strategy. In the radar-visual fusion autonomous driving device, the radar-visual fusion state vector corresponding to the initial autonomous driving strategy is compared with the associated descriptions in previous intelligent driving state assessment records to calculate the strategy's adaptability to the current road network quality characteristics. For example, the radar-visual fusion state vector corresponding to an initial autonomous driving strategy may indicate a narrow lidar detection range and weak visual sensor target recognition capabilities in low visibility conditions. If the current road network quality characteristics are heavy snow (extremely low visibility) and narrow roads (narrowed by heavy snow), the strategy's adaptability to these conditions will be low, making it unlikely to be adopted. Conversely, if a strategy's radar-visual fusion state vector indicates a wide detection range and strong target recognition capabilities in inclement weather, it will have a higher adaptability value under the same road network conditions and be more likely to be selected as the target autonomous driving strategy.

[0059] As can be seen, the embodiments of the present invention can generate multiple initial autonomous driving strategies based on driver request information. By mining the task elements representing each strategy using the radar-visual fusion state vector, the autonomous driving system can comprehensively and meticulously analyze the execution requirements of each strategy under radar-visual fusion technology. The intelligent driving state assessment records are obtained and used to calculate the road network environment quality characteristics for each initial strategy. This process fully utilizes historical empirical data, making the strategy evaluation more reliable and scientific. By selecting the target autonomous driving strategy and issuing it to the target electric snowmobile, the driving strategy most suitable for the current road network environment is selected. This overall improves the adaptability and safety of electric snowmobile autonomous driving. Regardless of road conditions (such as snow depth and road width), traffic flow (vehicle or skier density), and weather conditions (snowfall, visibility, wind speed), the vehicle can make optimal decisions, effectively avoiding driving risks caused by inappropriate strategy selection and improving the user's travel experience.

[0060] In some examples, each of the initial autonomous driving strategies is composed of one or more policy task nodes; mining a radar-visual fusion state vector for each of the initial autonomous driving strategies to obtain a radar-visual fusion state vector corresponding to each of the initial autonomous driving strategies includes: determining any initial autonomous driving strategy among the X initial autonomous driving strategies as a target initial autonomous driving strategy; obtaining a node depth of each of the policy task nodes in the target initial autonomous driving strategy, and prioritizing one or more policy task nodes in the target initial autonomous driving strategy based on the node depth of each policy task node to obtain a first policy task node chain; mining a node radar-visual fusion state vector for each of the policy task nodes in the first policy task node chain to obtain a node radar-visual fusion state vector corresponding to each of the policy task nodes; prioritizing the node radar-visual fusion state vectors corresponding to the one or more policy task nodes based on the priority of the one or more policy task nodes in the first policy task node chain to obtain a node radar-visual fusion state vector chain; and determining the node radar-visual fusion state vector chain as the radar-visual fusion state vector corresponding to the target initial autonomous driving strategy.

[0061] Furthermore, the node radar fusion state vector mining is performed on each of the policy task nodes in the first policy task node chain to obtain the node radar fusion state vector corresponding to each of the policy task nodes, including: determining any policy task node in the first policy task node chain as the target policy task node; performing multi-modal node element identification on the target policy task node to obtain the multi-modal node element vector of the target policy task node; performing mutual attention operation on the multi-modal node element vector of the target policy task node to obtain the node radar fusion state vector of the target policy task node.

[0062] Furthermore, the mutual attention operation is performed on the multi-mode node element vector of the target strategy task node to obtain the node radar fusion state vector of the target strategy task node, including: based on the task node conduction characteristics of the target strategy task node, a feature connection operation is performed on the multi-mode node element vector of the target strategy task node to obtain the task connection characteristics of the target strategy task node; according to the radar weight enhancement rule, the task connection characteristics of the target strategy task node are subjected to radar weight enhancement to obtain the task connection weighted characteristics of the target strategy task node; and the task connection weighted characteristics of the target strategy task node are determined as the node radar fusion state vector of the target strategy task node.

[0063] In actual application, the radar-visual fusion state vectors of each of the initial autonomous driving strategies are mined, and the radar-visual fusion state vectors corresponding to each of the initial autonomous driving strategies are obtained. The detailed description is as follows.

[0064] 1. The overall process of initial autonomous driving strategy construction and radar-visual fusion state vector mining

[0065] 1) Structure of the initial autonomous driving strategy

[0066] In this embodiment of the present invention, each initial autonomous driving policy is composed of one or more policy task nodes. These policy task nodes are the basic units that constitute the initial autonomous driving policy. Each node carries specific task information, and these task information together constitute the entire autonomous driving policy. For example, an initial autonomous driving policy may include policy task nodes such as "Detect Road Conditions Ahead," "Adjust Speed," and "Maintain Safe Distance."

[0067] 2) Process of mining the state vector of the radar-visual fusion

[0068] 2.1) Determine the Target Initial Autonomous Driving Strategy: First, any of the X initial autonomous driving strategies is identified as the target initial autonomous driving strategy. This step is to mine the radar-visual fusion state vector for each initial autonomous driving strategy.

[0069] 2.2) Prioritization of strategy task nodes

[0070] Obtaining Node Depth: Obtain the node depth of each policy task node in the target initial autonomous driving policy. Node depth reflects the hierarchical relationship of policy task nodes within the overall policy. For example, "Detect Road Conditions Ahead" might be a relatively basic task node, with a shallow node depth of 1. Meanwhile, "Adjust Speed Based on Road Conditions Ahead" might rely on the results of the "Detect Road Conditions Ahead" node, so its node depth might be 2.

[0071] Prioritization: Prioritize one or more policy task nodes in the target initial autonomous driving policy based on their node depths, generating a first policy task node chain. Within this node chain, nodes are sorted from highest to lowest priority. For example, based on the previously set node depth, the "Detect Road Ahead" node would be prioritized before the "Adjust Speed Based on Road Ahead" node.

[0072] 2.3) Node-Ray Vision Fusion State Vector Mining

[0073] Single node operation: Perform node radar and visual fusion state vector mining on each strategy task node in the first strategy task node chain to obtain the node radar and visual fusion state vector corresponding to each strategy task node. This process is performed separately for each strategy task node.

[0074] Constructing a node-level radar-visual fusion state vector chain: Based on the priority of one or more policy task nodes in the first policy task node chain, prioritize the node-level radar-visual fusion state vectors corresponding to the one or more policy task nodes to obtain a node-level radar-visual fusion state vector chain. For example, if there are three policy task nodes A, B, and C, and their corresponding node-level radar-visual fusion state vectors are VA, VB, and VC, respectively, the node-level radar-visual fusion state vector chain can be [VA, VB, VC] in order of priority.

[0075] Determine the radar-visual fusion state vector corresponding to the target initial autonomous driving strategy: Finally, the node radar-visual fusion state vector chain is determined to be the radar-visual fusion state vector corresponding to the target initial autonomous driving strategy. This radar-visual fusion state vector comprehensively reflects the relevant states of each task node of the target initial autonomous driving strategy under the radar-visual fusion technology.

[0076] 2. Detailed steps for mining node radar-visual fusion state vectors

[0077] 1) Determination of target strategy task nodes

[0078] When mining the node radar-vision fusion state vector of each policy task node in the first policy task node chain, any policy task node in the first policy task node chain is determined as the target policy task node. This is to process each node step by step.

[0079] 2) Identification of multimodal node elements and acquisition of multimodal node element vectors

[0080] Multimodal node element identification is performed for the target strategy task node. For example, for the target strategy task node "Detect road conditions ahead," multimodal node elements may include the lidar's detection range (e.g., within 100 meters ahead), detection frequency (once every 0.2 seconds), the visual sensor's target type (such as vehicles, pedestrians, traffic signs), and image clarity requirements (720p). These elements together constitute the multimodal node element vector for the target strategy task node. For example, a vector representing this multimodal node element vector might be [100, 0.2, 111, 720], where 111 is a code indicating the recognition of three target types: vehicles, pedestrians, and traffic signs.

[0081] 3) Mutual attention operation to obtain the node radar fusion state vector

[0082] 3.1) Feature Concatenation: Based on the target policy task node's task node conduction feature, a feature concatenation operation is performed on the target policy task node's multimodal node element vector to obtain the target policy task node's task connection feature. The task node conduction feature can be a feature related to the information conduction of the task node within the entire policy. For example, a task node conduction feature is a weight vector [0.5, 0.3, 0.2], and a feature concatenation operation (such as a weighted summation operation) is performed on the multimodal node element vector [100, 0.2, 111, 720] to obtain the task connection feature.

[0083] 3.2) Radar Weight Enhancement: Based on the radar weight enhancement rule, the task connection features of the target strategy task node are enhanced with radar weights to obtain the target strategy task node's weighted task connection features. The radar weight enhancement rule may assign different weights based on the importance of lidar and vision sensors in the radar-visual fusion technology. For example, the weight of lidar-related factors may be increased by 0.1, while the weight of vision-related factors may be adjusted based on their importance. For example, after radar weight enhancement, the task connection weighted features become [110, 0.22, 120, 750] (these values are just examples; actual calculations will depend on the specific weight enhancement rule).

[0084] 3.3) Determining the Node Radar and Vision Fusion State Vector: The target strategy task node's task connection weighted characteristics are used to determine its node radar and vision fusion state vector. This node radar and vision fusion state vector accurately reflects the target strategy task node's state under radar and vision fusion technology, taking into account multiple factors such as multi-mode node elements, task node transmission characteristics, and radar and vision weight enhancement.

[0085] This design first decomposes the initial autonomous driving policy into policy task nodes and prioritizes them based on node depth. This allows for more detailed analysis of each policy's internal structure and execution order, improving the rationality of policy execution. For example, when handling complex road conditions, the high-priority "Detect Road Conditions Ahead" node is executed first, enabling timely acquisition of environmental information and enabling subsequent nodes to make accurate decisions. Second, through multimodal node element recognition and mutual attention operations, a node-based radar-visual fusion state vector is derived, which comprehensively and accurately reflects the state of each policy task node under radar-visual fusion technology. For a specific numerical example, the node-based radar-visual fusion state vector, derived through precise calculation and weighting under varying multimodal node elements such as detection range and detection frequency, provides a more reliable basis for evaluating the policy's adaptability in a radar-visual fusion environment. Finally, this overall process helps improve the accuracy and adaptability of autonomous driving policies, thereby enhancing the safety and efficiency of electric snowmobile autonomous driving.

[0086] In the next step, based on the task node conduction characteristics of the target policy task node, a feature connection operation is performed on the multi-mode node element vector of the target policy task node to obtain the task connection characteristics of the target policy task node, including: performing a node clustering operation on the target policy task node based on the task node conduction characteristics of the target policy task node to obtain the target node cluster to which the target policy task node belongs; updating the first intelligent driving control event feature in the multi-mode node element vector of the target policy task node to the second intelligent driving control event feature used to characterize the vehicle-mounted human-computer interaction event indicated by the target node cluster to obtain the task connection characteristics of the target policy task node.

[0087] Based on this, the task connection features of the target strategy task node are subjected to thunder vision weight enhancement rules to obtain the task connection weighted features of the target strategy task node, including: obtaining the connection features to be processed whose feature attention distribution does not match the thunder vision weight enhancement rules from the task connection features of the target strategy task node; adjusting the attention weight of the connection features to be processed based on the thunder vision weight enhancement rules to obtain the target connection features; the feature attention distribution of the target connection features matches the thunder vision weight enhancement rules; determining the feature set composed of the target connection features and the remaining connection features as the task connection weighted features of the target strategy task node; the remaining connection features refer to the connection features in the task connection features of the target strategy task node other than the connection features to be processed.

[0088] In an embodiment of the present invention, the specific implementation methods of "based on the task node conduction characteristics of the target strategy task node, performing a feature connection operation on the multi-mode node element vector of the target strategy task node to obtain the task connection characteristics of the target strategy task node" and "according to the thunder vision weight enhancement rule, performing thunder vision weight enhancement on the task connection characteristics of the target strategy task node to obtain the task connection weighted characteristics of the target strategy task node" are as follows.

[0089] 1. Feature connection operation based on task node conduction features

[0090] 1) Node clustering and determination of target node clusters

[0091] Based on the target policy task node's task node transmission characteristics, a feature connection operation is performed on the multimodal node element vectors of the target policy task node to obtain the task connection characteristics. The target policy task node is first clustered based on the task node transmission characteristics of the target policy task node. The task node transmission characteristics contain characteristics related to task node information transmission and interaction. These characteristics can be used to divide the target policy task nodes into specific groups, namely, target node clusters.

[0092] For example, task node transmission characteristics contain some identifying information, such as task type identifiers and information flow direction identifiers. If clustering is performed based on task type identifiers, a target strategy task node with a "road condition detection and processing" type identifier might be assigned to a specific node cluster. This node cluster might also contain other task nodes related to road conditions, such as "road sign recognition" and "road roughness detection." For example, this node cluster is labeled Cluster 1.

[0093] 2) Feature update of multi-mode node feature vectors

[0094] After obtaining the target node cluster to which the target strategy task node belongs, the first intelligent driving control event feature in the multi-mode node element vector of the target strategy task node is updated to the second intelligent driving control event feature used to characterize the in-vehicle human-computer interaction event indicated by the target node cluster, thereby obtaining the task connection feature of the target strategy task node.

[0095] Taking a specific numerical example, for example, the multimodal node element vector of the target policy task node is [100, 0.2, 111, 720, 5], where the last value 5 represents the first intelligent driving control event feature, which may represent a specific driving control mode, such as manual driving assistance mode. If the vehicle-mounted human-computer interaction event indicated by the target node cluster (cluster 1) corresponds to the fully autonomous driving mode, then the second intelligent driving control event feature can be 1, indicating the fully autonomous driving mode. The updated multimodal node element vector (i.e., the task connection feature) becomes [100, 0.2, 111, 720, 1]. This update makes the multimodal node element vector of the target policy task node match the vehicle-mounted human-computer interaction event of the node cluster to which it belongs, thereby better reflecting the task connection characteristics in a specific cluster environment.

[0096] 2. Weight enhancement operation based on the rules of thunder vision weight enhancement

[0097] 1) Acquisition of pending connection features

[0098] In the process of performing thunder-visual weight enhancement on the task connection features of the target strategy task node according to the thunder-visual weight enhancement rule to obtain the task connection weighted features, the connection features to be processed whose feature attention distribution does not match the thunder-visual weight enhancement rule are first obtained from the task connection features of the target strategy task node.

[0099] For example, the radar weight enhancement rule stipulates that lidar-related factors (such as detection range and detection frequency) should account for 40% of the total attention weight of a specific task connection feature, visual sensor-related factors (such as target type recognition and image clarity requirements) should account for 30%, and other related factors (such as intelligent driving control event features) should account for 30%. If the task connection feature is [100, 0.2, 111, 720, 1], and analysis shows that the total attention weight of lidar-related factors is only 30%, then these lidar-related factors (such as detection range 100 and detection frequency 0.2) may be identified as connection features to be processed.

[0100] 2) Attention weight adjustment and acquisition of target connection features

[0101] Based on the radar weight enhancement rule, the attention weights of the connection features to be processed are adjusted to obtain the target connection features. Continuing with the above example, for the lidar-related features to be processed, the radar weight enhancement rule requires increasing their total attention weight to 40%. For example, a weight adjustment algorithm can be used, such as increasing the weight of each feature by a certain ratio. For example, if the original weight for a detection range of 100 is 0.15, and the original weight for a detection frequency of 0.2 is 0.15, after adjustment, the weight for a detection range of 100 might become 0.2, and the weight for a detection frequency of 0.2 might become 0.2 (these values are just examples; actual calculations depend on the specific adjustment algorithm). The adjusted connection features (i.e., the target connection features) become [120, 0.25, 111, 720, 1] (the changes in the detection range and detection frequency here are to reflect the effect of the weight adjustment, not a simple numerical correspondence). At this point, the feature attention distribution of the target connection features matches the radar weight enhancement rule.

[0102] 3) Determination of task connection weighted features

[0103] The feature set consisting of the target connection feature and the remaining connection feature is determined as the task connection weighted feature of the target strategy task node. In the above example, the remaining connection feature is [111, 720, 1]. Therefore, the task connection weighted feature is the feature set consisting of [120, 0.25] in the target connection feature [120, 0.25, 111, 720, 1] and the remaining connection feature [111, 720, 1], that is, [120, 0.25, 111, 720, 1]. This task connection weighted feature not only satisfies the radar and vision weight enhancement rules but also comprehensively considers the weight adjustment of different factors, more accurately reflecting the characteristics of the target strategy task node under the radar and vision fusion technology.

[0104] With this design, firstly, through node clustering operations and feature updates of multi-mode node element vectors, the target strategy task nodes can be better matched with the on-board human-computer interaction events of the node cluster to which they belong, which helps to improve the adaptability of the autonomous driving strategy in different interaction scenarios. For example, accurately adjusting the intelligent driving control event characteristics according to the requirements of the node cluster can ensure the correct operation of the vehicle in different driving modes. Secondly, the weight enhancement operation performed according to the radar and vision weight enhancement rules obtains and adjusts the connection features to be processed so that the attention distribution of the task connection features meets the requirements of the rules. Taking the specific attention weight ratio requirements as an example, this operation can reasonably distribute the weights of factors such as lidar and visual sensors, improve the accuracy and effectiveness of radar and vision fusion technology in the autonomous driving strategy, and thus improve the performance and safety of the entire autonomous driving system.

[0105] Under some preferred design ideas, the processing of the obtained driving user request information to generate X initial autonomous driving strategies for the driving user request information includes: receiving the driving user request information uploaded by the vehicle-mounted server; performing intelligent driving demand mining on the driving user request information to obtain an intelligent driving demand feature map of the driving user request information; and mapping and decoding a strategy vector based on the intelligent driving demand feature map to generate X initial autonomous driving strategies for the driving user request information.

[0106] In the embodiment of the present invention, the technical solution described in the preferred design concept is described in detail as follows.

[0107] 1. Overall Process of Generating Initial Autonomous Driving Strategies

[0108] 1) Receiving driving user request information

[0109] Under this preferred design approach, the driver's request information uploaded by the vehicle server must first be received. As the core hub for information exchange within the vehicle, the vehicle server uploads driver request information that contains the user's various expectations and requirements for autonomous driving. For example, the driver's request information may be transmitted in a specific data format, including destination information (in the form of geographic coordinates, such as longitude 120.123 and latitude 30.456), the desired driving speed range (such as a maximum speed of no more than 60 km / h), and the required driving comfort (such as smooth driving, avoiding sudden acceleration and braking).

[0110] 2) Intelligent Driving Demand Mining and Acquisition of Intelligent Driving Demand Characteristic Maps

[0111] Intelligent Driving Demand Mining: This process involves conducting intelligent driving demand mining on received driver request information. This process involves an in-depth analysis of the various implicit requirements related to intelligent driving contained in the user request information. For example, if the user request information includes a destination of a mountain ski resort and the current weather is snowing, in addition to the clear destination coordinates, it also implies that the vehicle needs to have the ability to cope with snowy road conditions (such as anti-skid control and recognition of snow-covered roads) and low visibility (such as visual sensor adjustment in rime weather).

[0112] Construction of an intelligent driving demand profile: Through intelligent driving demand mining, an intelligent driving demand profile is generated based on driver request information. This profile is a comprehensive description of intelligent driving needs, expressed in graphical or vector form. For example, when constructing an intelligent driving demand profile in vector form, it may include values from multiple dimensions. For example, if the vector dimensions are [road condition adaptability, speed control requirements, safety requirements, and environmental awareness requirements], for a request to a mountain ski resort, the road condition adaptability dimension might correspond to a high value, such as 0.8 (between 0 and 1, with 0.8 indicating a high requirement for snowy road conditions). The speed control requirement dimension might correspond to a speed range, such as the coded value for [0, 60] km / h. The safety requirement dimension might be set to a high value, such as 0.7 (indicating a high requirement for safe distance and stable driving), based on the user's comfort requirements. The environmental awareness requirement dimension might be set to 0.9 due to snowy weather (indicating a high requirement for identifying objects in snowy conditions and low visibility).

[0113] 3) Mapping and decoding of strategy vectors and generation of initial autonomous driving strategy

[0114] Mapping and decoding of strategy vectors: Mapping and decoding of strategy vectors are performed based on the intelligent driving demand characteristic map. This process involves converting the various characteristic values in the intelligent driving demand characteristic map into specific autonomous driving strategy elements. For example, based on the value of 0.8 in the road condition adaptability dimension, it may be determined during the mapping and decoding process that the vehicle needs to have a specific snow driving mode, such as adjusting tire pressure, optimizing the power distribution of the four-wheel drive system, etc. For the speed control requirement dimension of [0, 60] km / h, it will be converted into a specific speed control strategy, such as dynamically adjusting the vehicle speed within this speed range according to road conditions and traffic rules on different road sections (such as flat snow roads, sloping snow slopes, etc.).

[0115] Generation of initial autonomous driving strategies: Through mapping and decoding of the strategy vectors, X initial autonomous driving strategies are ultimately generated based on the driver's request information. Here, X is a positive integer, and its specific value depends on various factors. For example, different initial autonomous driving strategies may be generated based on different route selections (such as the shortest path, the safest path, and the most scenic path), different vehicle performance adjustment strategies (such as adjustments in energy-saving mode and high-performance mode), and other factors. For example, based on the above user request information, three initial autonomous driving strategies are generated (X=3):

[0116] Strategy 1: Choose the shortest path to the destination in energy-saving mode. Dynamically adjust the speed within [0, 60] km / h based on road conditions. The vehicle is set to energy-saving snow driving mode (such as reducing engine power output and optimizing the energy consumption of the anti-skid system). A safe distance is set to 3 meters. The visual sensor and lidar operate in a mode adapted to the low visibility of snow (such as reducing the visual sensor frame rate and adjusting the lidar detection frequency).

[0117] Strategy 2: Choose the safest route to the destination, using standard performance mode, maintaining a speed of around 40 km / h, adjusting the vehicle to snow safety driving mode (such as enhancing anti-skid control and increasing safety distance warning sensitivity), setting the safety distance to 5 meters, and operating the visual sensor and lidar in high-precision mode (such as improving the resolution of the visual sensor and increasing the detection accuracy of the lidar).

[0118] Strategy 3: Choose the most scenic route to the destination, using high-performance mode. The speed is dynamically adjusted based on road conditions within the range of [0, 60] km / h. The vehicle is adjusted to high-performance snow driving mode (such as maximizing engine power output to cope with possible climbing conditions and optimizing the handling performance of the four-wheel drive system). The safety distance is set to 4 meters, and the visual sensor and lidar operate in a balanced mode (while ensuring a certain level of visual effect and detection accuracy, taking into account energy consumption and system resource usage).

[0119] Therefore, firstly, by mining the intelligent driving demand of driving user request information to obtain the intelligent driving demand characteristic map, we can fully and deeply understand the user's needs, and no longer be limited to the surface request content. Taking the numerical examples in the intelligent driving demand characteristic map as an example, the numerical values of different dimensions accurately reflect the degree of user demand in terms of road conditions, speed, safety and environmental perception. Secondly, based on the intelligent driving demand characteristic map, the strategy vector is mapped and decoded to generate the initial autonomous driving strategy, making the generated strategy more targeted and reasonable. For example, the differences in different initial autonomous driving strategies in terms of path selection, vehicle mode, speed control and sensor working mode can meet the diverse needs of users and adapt to different driving conditions, thereby improving the flexibility and adaptability of the autonomous driving system, thereby improving the user experience and the safety of autonomous driving.

[0120] Under some other optional design ideas, the road network environment quality characteristics of each initial autonomous driving strategy are calculated through the association description between the radar and vision fusion state vectors corresponding to each initial autonomous driving strategy, the past radar and vision fusion state vector set in the intelligent driving state evaluation record, and the past road network environment quality characteristic set, including: determining any initial autonomous driving strategy among the X initial autonomous driving strategies as the target initial autonomous driving strategy; roaming the past radar and vision fusion state vector set in the intelligent driving state evaluation record based on the radar and vision fusion state vector corresponding to the target initial autonomous driving strategy; if there is a radar and vision fusion state vector corresponding to the target initial autonomous driving strategy in the past radar and vision fusion state vector set, If there is a target past radar-vision fusion state vector that is identical to the radar-vision fusion state vector corresponding to the target past radar-vision fusion state vector, the past road network environment quality characteristics corresponding to the target past radar-vision fusion state vector are determined as the road network environment quality characteristics of the target initial autonomous driving strategy; if there is no target past radar-vision fusion state vector identical to the radar-vision fusion state vector corresponding to the target initial autonomous driving strategy in the past radar-vision fusion state vector set, the target initial autonomous driving strategy is converted to obtain Y autonomous driving sub-strategies, and based on the Y autonomous driving sub-strategies, the association description between the past radar-vision fusion state vector set in the intelligent driving status evaluation record and the past road network environment quality characteristic set, the road network environment quality characteristics of the target initial autonomous driving strategy are determined; Y is a positive integer.

[0121] In the next step, the road network environment quality characteristics of the target initial autonomous driving strategy are determined based on the Y autonomous driving sub-strategies, the association description between the past radar-visual fusion state vector set in the intelligent driving status evaluation record and the past road network environment quality characteristic set, including: obtaining the local radar-visual fusion state vector corresponding to each of the autonomous driving sub-strategies; wandering the past radar-visual fusion state vector set according to the local radar-visual fusion state vector corresponding to each of the autonomous driving sub-strategies; and determining the road network environment quality characteristics of the target initial autonomous driving strategy based on the wandering results of the past radar-visual fusion state vector set.

[0122] In a further embodiment, the target initial autonomous driving strategy is composed of one or more strategy task nodes; the wandering result of the past radar vision fusion state vector set is the result of the target local radar vision fusion state vector being included in the past radar vision fusion state vector set, and the target local radar vision fusion state vector refers to any local radar vision fusion state vector among Y local radar vision fusion state vectors; the road network environment quality characteristics of the target initial autonomous driving strategy are determined based on the wandering result of the past radar vision fusion state vector set, including: determining the autonomous driving secondary strategy corresponding to the target local radar vision fusion state vector as the target autonomous driving secondary strategy, and setting the target autonomous driving secondary strategy as the target autonomous driving secondary strategy. The policy task nodes slightly included are determined as policy task quality labeling nodes; the policy task nodes other than the policy task quality labeling nodes in the target initial autonomous driving strategy are determined as policy task nodes to be processed, and the radar sensor acquisition data corresponding to the policy task nodes to be processed are obtained; the radar sensor acquisition data includes the past road network environment quality characteristics corresponding to the target local radar fusion state vector; the node environment quality sub-characteristics corresponding to the policy task node to be processed are determined based on the radar sensor acquisition data corresponding to the policy task node to be processed; the road network environment quality characteristics of the target initial autonomous driving strategy are determined based on the node environment quality sub-characteristics corresponding to the policy task node to be processed.

[0123] It can be understood that the above optional design ideas and the subsequent two embodiments are described in detail as follows.

[0124] 1. Overall Process for Calculating Road Network Environment Quality Characteristics for Initial Autonomous Driving Strategies

[0125] (1) Determination of the initial autonomous driving strategy and the wandering of the radar-visual fusion state vector

[0126] 1) Determination of the target initial autonomous driving strategy

[0127] When calculating the road network quality characteristics for each initial autonomous driving strategy, one of the X initial autonomous driving strategies is first selected as the target initial autonomous driving strategy. This selection allows for the calculation of road network quality characteristics for each initial autonomous driving strategy. X, a positive integer, represents the number of initial autonomous driving strategies. For example, X = 5 means there are five different initial autonomous driving strategies to choose from. Each strategy has its own unique components, potentially involving different driving route planning, speed control strategies, safety distance settings, and other aspects.

[0128] 2) Walk through the set of previous fusion state vectors based on the fusion state vector of the radar

[0129] Once the target initial autonomous driving strategy is determined, it is necessary to navigate through the set of past radar-visual fusion state vectors in the intelligent driving state assessment record based on the radar-visual fusion state vector corresponding to the strategy. The radar-visual fusion state vector is a multi-dimensional vector that contains various state information related to the lidar and visual sensors. For example, for the radar-visual fusion state vector of the target initial autonomous driving strategy, the lidar-related dimensions may include detection range (for example, 120 meters), detection frequency (for example, once every 0.15 seconds), horizontal resolution (for example, 0.3 degrees), etc.; the visual sensor-related dimensions may include image clarity requirements (for example, 1080p), target type identification coding (for example, represented by binary coding, identifying vehicles, pedestrians, and traffic signs as 111 respectively), etc.

[0130] When navigating through the set of past radar-visual fusion state vectors, each dimension of the target initial autonomous driving policy's radar-visual fusion state vector is compared with each vector in the set of past radar-visual fusion state vectors. For example, a vector in the set of past radar-visual fusion state vectors might be [110 meters, every 0.18 seconds, 0.25 degrees, 720p, 101]. The two vectors need to be compared to see if their values in each dimension are identical or match within a certain error range.

[0131] (2) Processing when there are identical radar-visual fusion state vectors

[0132] 1) Determine the road network environmental quality characteristics

[0133] If a target past radar-visual fusion state vector is found in the set of past radar-visual fusion state vectors that is identical to the radar-visual fusion state vector corresponding to the target initial autonomous driving policy, the past road network environment quality characteristics corresponding to the target past radar-visual fusion state vector are directly determined as the road network environment quality characteristics of the target initial autonomous driving policy. Past road network environment quality characteristics are also a set of information containing multiple aspects. For example, they may include road snow depth (e.g., 20 cm), road flatness (expressed as slope, such as a 3% slope), traffic flow (8 vehicles per kilometer), and weather conditions (e.g., light snow, visibility of 400 meters). These past road network environment quality characteristics are directly mapped to the target initial autonomous driving policy as its road network quality characteristics, providing a direct basis for evaluating the policy's applicability in similar current road network environments.

[0134] (III) Processing when there is no identical radar-visual fusion state vector

[0135] 1) Conversion of target initial autonomous driving strategy

[0136] If the target past radar-visual fusion state vector does not exist in the set of past radar-visual fusion state vectors that is identical to the radar-visual fusion state vector corresponding to the target initial autonomous driving policy, the target initial autonomous driving policy needs to be transformed to obtain Y autonomous driving sub-policies. This transformation is a rule-based adjustment aimed at generating a policy that is more comparable to past experience. For example, if the target initial autonomous driving policy originally set a fixed speed of 50 km / h, the transformed autonomous driving sub-policy might dynamically adjust the speed between [45, 55] km / h (an example case when Y = 1). If Y = 3, there may be three different autonomous driving sub-policies, each adjusting the target initial autonomous driving policy from different perspectives, such as changing the driving route selection, the basis for speed adjustment, or the safety distance setting.

[0137] 2) Determine road network environmental quality characteristics based on autonomous driving secondary strategies

[0138] Based on the Y autonomous driving secondary strategies, the association description between the past radar-visual fusion state vector set in the intelligent driving state evaluation record and the past road network environment quality feature set, the road network environment quality features of the target initial autonomous driving strategy are determined. This step involves the following sub-steps:

[0139] 2. Steps to Determine Road Network Environmental Quality Characteristics Based on Autonomous Driving Sub-Strategies

[0140] (1) Obtain the local radar-visual fusion state vector and walk through the set of past radar-visual fusion state vectors

[0141] 1) Acquisition of local radar-visual fusion state vector

[0142] First, we need to obtain the local radar-visual fusion state vectors corresponding to each autonomous driving sub-strategy. These local radar-visual fusion state vectors are specifically constructed for autonomous driving sub-strategies. Their structure and meaning are similar to the radar-visual fusion state vectors, but they reflect the characteristics of the autonomous driving sub-strategy. For example, for an autonomous driving sub-strategy, the lidar detection range in its local radar-visual fusion state vector may be changed to 110 meters (adjusted from 120 meters in the target initial autonomous driving strategy), the detection frequency may be changed to once every 0.12 seconds, and the visual sensor image resolution requirement may be changed to 720p (adjusted from 1080p).

[0143] 2) Walk through the set of past radar fusion state vectors based on the local radar fusion state vector

[0144] Based on the local radar-visual fusion state vectors corresponding to each autonomous driving secondary policy, a walk operation is performed again within the set of past radar-visual fusion state vectors. Similar to the walk operation based on the target initial autonomous driving policy, each dimension of the local radar-visual fusion state vector is compared with vectors in the set of past radar-visual fusion state vectors to find a matching vector.

[0145] 2. Determining the road network environment quality characteristics of the target initial autonomous driving strategy based on the wandering results

[0146] 1) Determination of target autonomous driving secondary strategy and strategy task quality annotation nodes

[0147] When the target initial autonomous driving policy consists of one or more policy task nodes, the walk result of the set of past radar-visual fusion state vectors is the result of including the target local radar-visual fusion state vector in the set of past radar-visual fusion state vectors. The target local radar-visual fusion state vector is any one of the Y local radar-visual fusion state vectors. Once such a target local radar-visual fusion state vector is found, the corresponding autonomous driving sub-policy is determined as the target autonomous driving sub-policy. For example, if, when Y = 3, the autonomous driving sub-policy corresponding to one of the local radar-visual fusion state vectors satisfies the walk result, then this autonomous driving sub-policy is determined as the target autonomous driving sub-policy.

[0148] Next, the policy task nodes included in the target autonomous driving sub-policy are identified as policy task quality-labeled nodes. For example, if the target autonomous driving sub-policy includes the three policy task nodes of "detecting road conditions ahead," "adjusting speed based on road conditions," and "maintaining a safe distance," then these three nodes are identified as policy task quality-labeled nodes.

[0149] 2) Obtaining the pending strategic task nodes and radar sensor data

[0150] Determine the policy task nodes in the target initial autonomous driving policy, except for the policy task quality labeling node, as pending policy task nodes. For example, if the target initial autonomous driving policy has a "Road Sign Recognition" node in addition to the three labeling nodes mentioned above, then "Road Sign Recognition" is a pending policy task node.

[0151] Obtain the radar sensor data corresponding to the pending strategy task node. This radar sensor data includes the road network environment quality characteristics corresponding to the target local radar fusion state vector. For example, for the pending strategy task node "Identify Road Signs," the road network environment quality characteristics in the radar sensor data might include the density of traffic signs on the road (5 traffic signs per kilometer) and the degree to which road sign clarity is affected by weather (e.g., light snow reduces sign clarity by 30%).

[0152] 3) Determination of node environmental quality sub-characteristics and road network environmental quality characteristics

[0153] The node environmental quality sub-feature corresponding to the pending strategic task node is determined based on the radar sensor data collected for that node. For example, for the pending strategic task node "Identify Road Signs," its node environmental quality sub-feature can be determined based on the density of traffic signs on the road and the degree to which sign clarity is affected by weather. If the traffic sign density is high and the sign clarity is significantly affected by weather, then this node environmental quality sub-feature might indicate a high difficulty level for identifying road signs in this road network environment. This could be expressed as a numerical value of 0.7 (where 0 represents easy and 1 represents very difficult).

[0154] Finally, the road network environmental quality characteristics of the target initial autonomous driving strategy are determined based on the node environmental quality sub-characteristics corresponding to the pending strategy task nodes. This may involve a comprehensive calculation algorithm. For example, if the target initial autonomous driving strategy has multiple pending strategy task nodes, and their node environmental quality sub-characteristics are calculated to be 0.7, 0.6, and 0.8, respectively, a weighted average algorithm (for example, with weights of 0.3, 0.3, and 0.4, respectively) is used to calculate the road network environmental quality characteristics of the target initial autonomous driving strategy to be (0.7 × 0.3 + 0.6 × 0.3 + 0.8 × 0.4) = 0.71.

[0155] Applying the above technical solution, first, matching vectors are found by traversing the set of past radar-visual fusion state vectors in the intelligent driving state assessment record, fully utilizing past experience data. When identical vectors exist, the road network environment quality characteristics are directly determined, which is efficient and accurate. Secondly, when identical vectors do not exist, the system converts to an autonomous driving secondary strategy and performs subsequent operations, increasing the flexibility of strategy evaluation. The process of determining strategy task quality labeling nodes and pending strategy task nodes, obtaining radar-visual sensor data to determine node environment quality sub-features, and then determining road network environment quality characteristics enables detailed analysis of each strategy task node under different road network environments, thereby improving the adaptability of the entire autonomous driving strategy to the road network environment and ensuring the safety and effectiveness of autonomous driving.

[0156] In some independent technical solutions, the target initial autonomous driving strategy is composed of one or more policy task nodes; the wandering result of the past radar-vision fusion state vector set is the result that the past radar-vision fusion state vector set does not include any local radar-vision fusion state vector; the road network environment quality characteristics of the target initial autonomous driving strategy are determined based on the wandering result of the past radar-vision fusion state vector set, including: obtaining the node depth of each policy task node in the target initial autonomous driving strategy, and prioritizing one or more policy task nodes in the target initial autonomous driving strategy based on the node depth of each policy task node to obtain a second policy task node chain corresponding to the target initial autonomous driving strategy; based on the priority order of the one or more policy task nodes in the second policy task node chain, sequentially mining the node environment quality sub-features corresponding to each policy task node; determining the node environment quality sub-features corresponding to the policy task end node as the road network environment quality characteristics of the target initial autonomous driving strategy; the policy task end node refers to the policy task node at the end of the node chain in the second policy task node chain.

[0157] On the basis of the above content, based on the wandering results of the past radar-visual fusion state vector set, the specific implementation of determining the road network environment quality characteristics of the target initial autonomous driving strategy is as follows.

[0158] 1. Overall process of determining the road network environment quality characteristics of the target initial autonomous driving strategy based on specific walk results

[0159] (1) Structure of the target initial autonomous driving strategy and strategy task nodes

[0160] 1) Composition of the target initial autonomous driving strategy

[0161] In this standalone technical solution, the target initial autonomous driving strategy consists of one or more policy task nodes. These policy task nodes are the fundamental building blocks of the target initial autonomous driving strategy. Each node performs a specific task or function, and they work together to implement the overall autonomous driving strategy. For example, a target initial autonomous driving strategy might include policy task nodes such as "initiating vehicle self-check," "planning the initial driving route," "detecting surrounding traffic participants," and "adjusting vehicle speed based on road conditions." These nodes are interconnected according to a specific logical sequence and dependencies, collectively determining the execution process and effectiveness of the autonomous driving strategy.

[0162] (2) Prioritization Based on Node Depth and Formation of the Second Strategy Task Node Chain

[0163] 1) Obtaining and Significance of Node Depth

[0164] First, we need to obtain the node depth of each policy task node in the target initial autonomous driving policy. Node depth reflects the hierarchical relationship and execution order of policy task nodes within the entire policy. For example, "initiating a vehicle self-check" might be the most basic operation, with a node depth of 1; "planning an initial driving route" might rely on the results of the vehicle self-check, with a node depth of 2; "detecting surrounding traffic participants" relies on route planning, with a node depth of 3; and "adjusting vehicle speed based on road conditions" relies on detecting surrounding traffic participants, with a node depth of 4.

[0165] 2) Priority Arrangement and Construction of the Second Strategy Task Node Chain

[0166] Based on the node depth of each policy task node, one or more policy task nodes in the target initial autonomous driving policy are prioritized to obtain a second policy task node chain corresponding to the target initial autonomous driving policy. In this node chain, the policy task nodes are arranged from high to low priority. For example, based on the node depth mentioned above, the second policy task node chain might be ["Initiate vehicle self-test," "Plan initial driving route," "Detect surrounding traffic participants," "Adjust vehicle speed based on road conditions"]. This priority order, determined by node depth, reflects the logical sequence of autonomous driving policy execution: basic, prerequisite tasks are executed first, followed by subsequent tasks that depend on the results of the previous tasks.

[0167] (III) Sequentially mining node environmental quality sub-features and determining road network environmental quality characteristics

[0168] 1) Mining node environmental quality sub-features sequentially

[0169] Based on the priority of one or more policy task nodes in the second policy task node chain, the node environmental quality sub-features corresponding to each policy task node are mined in order. For each policy task node, its node environmental quality sub-feature reflects the difficulty of the node in performing the task in a specific road network environment or the related environmental quality influencing factors. For example, for the "Start Vehicle Self-Inspection" node, its node environmental quality sub-feature may be related to environmental factors such as the temperature and humidity at the vehicle's location. For example, if the temperature at the vehicle's location is -10°C, based on empirical data, there may be certain risks in vehicle self-inspection in such a low temperature environment. Its node environmental quality sub-feature can be represented by a numerical value, such as 0.6 (0 indicates no risk and 1 indicates high risk).

[0170] For the "Plan Initial Driving Route" node, its node environmental quality sub-characteristic may be related to road network congestion, road construction conditions, and other factors. If the current road network congestion index is 0.4 (0 indicates no congestion, 1 indicates severe congestion), and the proportion of road construction sections is 0.1 (i.e., 10% of the road sections are under construction), using a specific algorithm (e.g., a weighted average algorithm, where the congestion index is weighted 0.6 and the proportion of road construction sections is weighted 0.4), the node environmental quality sub-characteristic for this node is calculated to be (0.4 × 0.6 + 0.1 × 0.4) = 0.28.

[0171] In this manner, the node environmental quality sub-features for nodes such as "Detect surrounding traffic participants" and "Adjust vehicle speed based on road conditions" are calculated sequentially. For example, the node environmental quality sub-feature for the "Detect surrounding traffic participants" node may be affected by factors such as weather conditions (such as visibility) and the density of traffic participants. For example, if visibility is 500 meters (lower visibility, greater detection difficulty) and the density of traffic participants is 30 per square kilometer (higher density, greater detection difficulty), a specific function (such as an inversely proportional function relationship between visibility and traffic participant density, followed by normalization) is used to obtain the node environmental quality sub-feature of 0.7 for this node.

[0172] 2) Determine the road network environment quality characteristics of the target initial autonomous driving strategy

[0173] The node environmental quality sub-feature corresponding to the end node of the strategy task is determined as the road network environmental quality feature of the target initial autonomous driving strategy. In the previously constructed second strategy task node chain ["Initiate Vehicle Self-Test," "Plan Initial Driving Route," "Detect Surrounding Traffic Participants," and "Adjust Speed Based on Road Conditions"], "Adjust Speed Based on Road Conditions" is the end node of the strategy task. For example, if the node environmental quality sub-feature of this node is 0.7, then this value 0.7 is determined as the road network environmental quality feature of the target initial autonomous driving strategy. This determination method reflects that the ultimate execution of the entire autonomous driving strategy depends largely on the execution of the last key task node in the road network environment. Using this as a representative road network environmental quality feature of the entire strategy can concisely and effectively reflect the adaptability and feasibility of the entire strategy in a specific road network environment.

[0174] This design, by obtaining the node depth of the policy task node and sorting it by priority to obtain the second policy task node chain, can clearly reflect the logical execution order of the target initial autonomous driving strategy. This helps to deeply understand the dependencies between various task nodes, thereby better evaluating the rationality of the entire strategy. Secondly, the node environmental quality sub-features are sequentially mined and the sub-features of the policy task terminal node are determined as the road network environmental quality features, considering the impact of each link in the entire policy execution process on the final result. Taking a specific numerical example, the sub-feature values calculated by different nodes based on their respective relevant environmental factors can accurately reflect the status of the node in the road network environment. The final determined road network environmental quality features can effectively measure the adaptability of the target initial autonomous driving strategy in a specific road network environment, help to screen out autonomous driving strategies that are more suitable for the current environment, and improve the safety and effectiveness of autonomous driving.

[0175] In other independent technical solutions, a target autonomous driving strategy is selected from the X initial autonomous driving strategies based on the road network environment quality characteristics of each of the initial autonomous driving strategies, including: obtaining a road network environment quality characteristic with the highest feature normalization value from the road network environment quality characteristics of the X initial autonomous driving strategies; and determining the initial autonomous driving strategy indicated by the road network environment quality characteristic with the highest feature normalization value as the target autonomous driving strategy for which the driving user requests information.

[0176] In this technical solution, the screening process of the target autonomous driving strategy is as follows.

[0177] 1. Overall process of selecting target autonomous driving strategies based on road network environment quality characteristics

[0178] (1) Obtaining feature normalization values

[0179] 1) Review of road network environmental quality characteristics

[0180] In previous technical solutions, road network quality characteristics were calculated for each initial autonomous driving strategy. These characteristics were derived by comprehensively considering various factors, such as road conditions (such as snow depth and smoothness), traffic flow (number of vehicles and pedestrians), and weather conditions (visibility, snowfall), all of which influence the execution of the autonomous driving strategy. Each initial autonomous driving strategy has its own corresponding road network quality characteristics, which are represented in a specific form, such as a vector containing multiple values or a value processed by a specific algorithm.

[0181] 2) Necessity and operation of feature normalization

[0182] To facilitate comparison of the road network environmental quality characteristics of different initial autonomous driving strategies, feature normalization is required. Feature normalization is the process of converting values of different ranges and dimensions to a unified range. For example, if the road network environmental quality characteristics of an initial autonomous driving strategy include the influencing values of three factors: snow depth, traffic flow, and visibility, the snow depth may be an actual value in centimeters (e.g., 15 centimeters), traffic flow is expressed in the number of vehicles per kilometer (e.g., 8 vehicles / kilometer), and visibility is expressed in meters (e.g., 400 meters). Feature normalization converts these values of different dimensions into values between 0 and 1.

[0183] Specific normalization algorithms can use different methods depending on the actual situation. Taking linear normalization as an example, if the maximum snow depth is set to 50 cm (this is set based on the maximum snow depth that may be encountered in practice), then the normalized value of a snow depth of 15 cm is 15 / 50 = 0.3. For traffic flow, for example, if the maximum traffic flow is set to 20 vehicles per kilometer, a traffic flow of 8 vehicles / kilometer is normalized to 8 / 20 = 0.4. For visibility, if the minimum visibility is set to 100 meters and the maximum visibility is set to 1000 meters, the normalized value of 400 meters of visibility is (400-100) / (1000-100) = 0.33. These normalized values are recombined or further processed to obtain the characteristic normalized values of the road network environmental quality characteristics of the initial autonomous driving strategy.

[0184] 2. Determination of the target autonomous driving strategy

[0185] 1) Obtain the road network environmental quality feature with the highest feature normalization value

[0186] After obtaining the normalized values of the road network environment quality characteristics for X initial autonomous driving strategies, the road network environment quality characteristic with the highest normalized value is selected from these values. For example, if there are three initial autonomous driving strategies (X = 3), and their normalized road network environment quality characteristics are 0.55, 0.6, and 0.45, respectively, then the characteristic with the highest normalized value is 0.6.

[0187] 2) Determine the target autonomous driving strategy

[0188] The initial autonomous driving strategy indicated by the road network quality characteristic with the highest feature normalization value is determined as the target autonomous driving strategy for the driver's requested information. Continuing with the above example, the initial autonomous driving strategy corresponding to a feature normalization value of 0.6 is determined as the target autonomous driving strategy. This determination is based on the principle that a higher feature normalization value indicates that the initial autonomous driving strategy is more adaptable to the current road network environment. Because this feature normalization value is derived by comprehensively considering various factors affecting autonomous driving (such as the aforementioned road conditions, traffic flow, and weather conditions), a higher value indicates that the strategy has a relative advantage in coping with the various conditions of the current road network environment.

[0189] This design, first, converts the road network environmental quality characteristics of different initial autonomous driving strategies into a unified measurement standard through feature normalization, solving the problem of difficulty in directly comparing numerical values of different dimensions and ranges. Taking a specific numerical example, different factors such as snow depth, traffic flow, and visibility can be compared on the same scale after normalization. Secondly, the target autonomous driving strategy is determined based on the highest feature normalization value, which can select the strategy that is most adaptable to the current road network environment. This helps to improve the efficiency and safety of autonomous driving, because the selected target autonomous driving strategy has relatively optimal performance when considering various environmental factors, thereby being able to better cope with actual driving conditions and reduce the risks that may be caused by improper strategy selection.

[0190] Furthermore, Figure 2 This is a schematic diagram of the structure of a radar-visual fusion automatic driving device 200 provided in an embodiment of the present invention. Figure 2 The radar-visual fusion automatic driving device 200 shown includes a processor 210, which can call and run a computer program from a memory to implement the method in the embodiment of the present invention.

[0191] Alternatively, as Figure 2 As shown, the radar-visual fusion automatic driving device 200 may further include a memory 230. The processor 210 may call and run a computer program from the memory 230 to implement the method in the embodiment of the present invention.

[0192] The memory 230 may be a separate device independent of the processor 210 , or may be integrated into the processor 210 .

[0193] Alternatively, as Figure 2 As shown, the radar-vision fusion automatic driving device 200 may further include a transceiver 220, and the processor 210 may control the transceiver 220 to interact with other devices. Specifically, it may send information or data to other devices, or receive information or data sent by other devices.

[0194] Optionally, the radar-vision fusion autonomous driving device 200 can implement the corresponding processes corresponding to the storage engine or components in the storage engine (such as a processing module) or the device deployed with the storage engine in each method of the embodiments of the present invention. For the sake of brevity, they will not be repeated here.

[0195] It should be understood that the processor in the embodiment of the present invention may be an integrated circuit chip with signal processing capabilities.

[0196] It is understood that the memory in the embodiments of the present invention may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. It should be noted that the memory of the systems and methods described herein is intended to include but is not limited to suitable types of memory.

[0197] Based on the above, a readable storage medium is provided, on which a program or instruction is stored. When the program or instruction is executed by a processor, the steps of the above method are implemented.

[0198] Based on the same or similar inventive concept as above, an electric snowmobile is also provided. The snowmobile can adopt a wheeled structure and can adopt an EMB brake as a braking system. The electric snowmobile is communicatively connected to a radar-vision fusion automatic driving device. The electric snowmobile is used to receive a target automatic driving strategy issued by the radar-vision fusion automatic driving device, and the target automatic driving strategy is determined by the above method.

[0199] It should be noted that, in this article, the terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device comprising a series of elements includes not only those elements, but also includes other elements that are not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element. In addition, it should be pointed out that the scope of the methods and devices in the embodiments of the present invention is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in the opposite order according to the functions involved. For example, the described method may be performed in an order different from that described, and various steps may also be added, omitted, or combined. In addition, the features described with reference to certain examples may be combined in other examples.

[0200] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solutions of the embodiments of the present invention, or the part that contributes to the prior art, can be embodied in the form of a computer software product. The computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for enabling a terminal (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0201] The above describes an embodiment of the present invention in conjunction with the accompanying drawings, but the embodiment of the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the embodiment of the present invention, ordinary technicians in this field can also make many forms without departing from the purpose of the embodiment of the present invention and the scope of protection of the embodiment of the present invention, all of which are protected by the embodiment of the present invention.

Claims

1. An autonomous driving method based on radar and vision fusion, characterized in that: The method comprises: Processing the acquired driving user request information to generate X initial autonomous driving strategies based on the driving user request information; X is a positive integer; Mining a radar-visual fusion state vector for each of the initial autonomous driving strategies to obtain a radar-visual fusion state vector corresponding to each of the initial autonomous driving strategies; a radar-visual fusion state vector corresponding to each of the initial autonomous driving strategies is used to represent a strategy execution task element of the initial autonomous driving strategy; Obtaining an intelligent driving status assessment record; the intelligent driving status assessment record includes a description of an association between a past set of radar-visual fusion state vectors and a past set of road network environment quality characteristics, wherein a past radar-visual fusion state vector in the past set of radar-visual fusion state vectors is associated with a past road network environment quality characteristic in the past set of road network environment quality characteristics, and a past radar-visual fusion state vector is a radar-visual fusion state vector corresponding to a past autonomous driving strategy for which a driving user requested information; Calculating the road network environment quality characteristics of each of the initial autonomous driving strategies by using the association description between the radar-visual fusion state vector corresponding to each of the initial autonomous driving strategies, the past radar-visual fusion state vector set in the intelligent driving state evaluation record, and the past road network environment quality characteristic set; A target autonomous driving strategy is selected from the X initial autonomous driving strategies based on the road network environment quality characteristics of each of the initial autonomous driving strategies, and the target autonomous driving strategy is sent to the target electric snowmobile.

2. The method according to claim 1, wherein Each of the initial autonomous driving strategies is composed of one or more strategy task nodes; The performing of radar-visual fusion state vector mining on each of the initial autonomous driving strategies to obtain the radar-visual fusion state vector corresponding to each of the initial autonomous driving strategies includes: Determining any one of the X initial autonomous driving strategies as a target initial autonomous driving strategy; Obtaining a node depth of each of the policy task nodes in the target initial autonomous driving strategy, and prioritizing one or more policy task nodes in the target initial autonomous driving strategy based on the node depth of each of the policy task nodes to obtain a first policy task node chain; Performing node radar-visual fusion state vector mining on each of the strategy task nodes in the first strategy task node chain to obtain a node radar-visual fusion state vector corresponding to each of the strategy task nodes; Based on the priority order of the one or more policy task nodes in the first policy task node chain, the node radar and vision fusion state vectors corresponding to the one or more policy task nodes are prioritized to obtain a node radar and vision fusion state vector chain; The node radar-visual fusion state vector chain is determined as the radar-visual fusion state vector corresponding to the target initial autonomous driving strategy.

3. The method according to claim 2, wherein The performing of node radar-visual fusion state vector mining on each of the policy task nodes in the first policy task node chain to obtain the node radar-visual fusion state vector corresponding to each of the policy task nodes includes: Determine any policy task node in the first policy task node chain as a target policy task node; Performing multi-mode node element identification on the target strategy task node to obtain a multi-mode node element vector of the target strategy task node; A mutual attention operation is performed on the multi-mode node element vector of the target strategy task node to obtain a node radar-visual fusion state vector of the target strategy task node.

4. The method according to claim 3, wherein The performing of a mutual attention operation on the multi-mode node element vector of the target strategy task node to obtain a node radar-visual fusion state vector of the target strategy task node includes: Based on the task node conduction feature of the target strategy task node, a feature connection operation is performed on the multi-mode node element vector of the target strategy task node to obtain the task connection feature of the target strategy task node; According to the thunder and light weight strengthening rule, the task connection feature of the target strategy task node is subjected to thunder and light weight strengthening to obtain the task connection weighted feature of the target strategy task node; The task connection weighted feature of the target strategy task node is determined as the node radar-vision fusion state vector of the target strategy task node.

5. The method according to claim 4, wherein The step of performing a feature connection operation on the multi-mode node element vector of the target policy task node based on the task node conduction feature of the target policy task node to obtain the task connection feature of the target policy task node includes: Performing a node clustering operation on the target policy task node based on the task node conduction characteristics of the target policy task node to obtain a target node cluster to which the target policy task node belongs; The first intelligent driving control event feature in the multi-mode node element vector of the target strategy task node is updated to the second intelligent driving control event feature used to characterize the in-vehicle human-computer interaction event indicated by the target node cluster, and the task connection feature of the target strategy task node is obtained.

6. The method according to claim 4, wherein The step of performing the thunder-visual weight enhancement on the task connection feature of the target strategy task node according to the thunder-visual weight enhancement rule to obtain the task connection weighted feature of the target strategy task node includes: Obtaining, from the task connection features of the target strategy task node, connection features to be processed whose feature attention distribution does not match the thunder vision weight reinforcement rule; Adjusting the attention weight of the connection feature to be processed based on the thunder-vision weight enhancement rule to obtain a target connection feature; the feature attention distribution of the target connection feature matches the thunder-vision weight enhancement rule; The feature set consisting of the target connection feature and the remaining connection features is determined as the task connection weighted feature of the target strategy task node; the remaining connection feature refers to the connection features in the task connection features of the target strategy task node excluding the connection features to be processed.

7. The method according to claim 1, wherein The processing of the acquired driving user request information to generate X initial autonomous driving strategies based on the driving user request information includes: Receive driver request information uploaded by the vehicle server; Performing intelligent driving demand mining on the driving user request information to obtain an intelligent driving demand feature map of the driving user request information; Mapping and decoding the strategy vector are performed based on the intelligent driving demand characteristic graph to generate X initial autonomous driving strategies based on the driving user request information.

8. The method according to claim 1, wherein Calculating the road network environment quality characteristics of each of the initial autonomous driving strategies by using the radar-visual fusion state vectors corresponding to each of the initial autonomous driving strategies, the association description between the past radar-visual fusion state vector set in the intelligent driving state evaluation record, and the past road network environment quality characteristic set includes: Determining any one of the X initial autonomous driving strategies as a target initial autonomous driving strategy; According to the radar-visual fusion state vector corresponding to the target initial autonomous driving strategy, the set of past radar-visual fusion state vectors in the intelligent driving state evaluation record is walked; If there is a target past radar-visual fusion state vector in the set of past radar-visual fusion state vectors that is identical to the radar-visual fusion state vector corresponding to the target initial autonomous driving strategy, determining the past road network environment quality feature corresponding to the target past radar-visual fusion state vector as the road network environment quality feature of the target initial autonomous driving strategy; If there is no target past radar-visual fusion state vector identical to the radar-visual fusion state vector corresponding to the target initial autonomous driving strategy in the past radar-visual fusion state vector set, converting the target initial autonomous driving strategy to obtain Y autonomous driving sub-strategies, and determining the road network environment quality characteristics of the target initial autonomous driving strategy based on the Y autonomous driving sub-strategies, the past radar-visual fusion state vector set in the intelligent driving state assessment record, and the past road network environment quality characteristic set; Y is a positive integer; The determining of the road network environment quality characteristics of the target initial autonomous driving strategy based on the Y autonomous driving secondary strategies, the association description between the past radar-visual fusion state vector set in the intelligent driving state assessment record, and the past road network environment quality characteristic set includes: Obtaining a local radar-visual fusion state vector corresponding to each of the autonomous driving secondary strategies; Walking the set of past radar-visual fusion state vectors according to the local radar-visual fusion state vectors corresponding to each of the autonomous driving secondary strategies; Determining the road network environment quality characteristics of the target initial autonomous driving strategy based on the wandering results of the past radar-visual fusion state vector set; The target initial autonomous driving strategy is composed of one or more strategy task nodes; the wandering result of the past radar-visual fusion state vector set is the result of the past radar-visual fusion state vector set including the target local radar-visual fusion state vector, and the target local radar-visual fusion state vector refers to any local radar-visual fusion state vector among the Y local radar-visual fusion state vectors; The determining of the road network environment quality characteristics of the target initial autonomous driving strategy based on the wandering results of the past radar-visual fusion state vector set includes: Determining the autonomous driving secondary strategy corresponding to the target local radar-visual fusion state vector as the target autonomous driving secondary strategy, and determining the strategy task node included in the target autonomous driving secondary strategy as the strategy task quality labeling node; Determine the strategy task nodes other than the strategy task quality labeling nodes in the target initial autonomous driving strategy as pending strategy task nodes, and obtain radar vision sensor acquisition data corresponding to the pending strategy task nodes; the radar vision sensor acquisition data includes past road network environment quality characteristics corresponding to the target local radar vision fusion state vector; Determining the node environment quality sub-feature corresponding to the strategy task node to be processed based on the radar sensor data corresponding to the strategy task node to be processed; The road network environment quality characteristics of the target initial autonomous driving strategy are determined based on the node environment quality sub-characteristics corresponding to the to-be-processed strategy task node.

9. A radar-visual fusion automatic driving device, characterized in that: The method comprises at least one processor and a memory; the memory stores computer-executable instructions; the at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor executes the method according to any one of claims 1 to 8.

10. An electric snowmobile, characterized in that: The electric snowmobile is communicatively connected to a radar-vision fusion automatic driving device, and the electric snowmobile is used to receive a target automatic driving strategy issued by the radar-vision fusion automatic driving device, wherein the target automatic driving strategy is determined by the method described in any one of claims 1-8.

Images