Intelligent path planning method and system for deicing vehicle chassis

Abstract

The application discloses a deicing vehicle chassis intelligent path planning method and system, and realizes automatic driving of the chassis of the deicing vehicle in a deicing process. The method comprises the following steps: step 1, loading a high-precision vector map of an airport; step 2, converting the high-precision vector map into a global road network topology map; step 3, performing path searching based on the global road network topology map; step 4, acquiring sensor data and identifying dynamic obstacles; step 5, performing local tracking and safety evaluation; step 6, performing dynamic obstacle avoidance in a lane; step 7, when the deicing vehicle reaches a target airport apron intersection, the deicing vehicle leaves a fixed service lane and enters a semi-open area around the apron, and is switched to a local cost map; step 8, generating a virtual safety envelope line according to an aircraft model; and step 9, using a local path planning algorithm and combining the virtual safety envelope line to enable the deicing vehicle to perform multi-degree-of-freedom fine adjustment to approach the aircraft body.

Description

Technical Field

[0001] This application relates to the field of autonomous driving and aviation ground service support equipment technology, and more specifically, to an intelligent path planning method and system for a de-icing vehicle chassis based on fixed airport routes. Background Technology

[0002] In the air transport sector, aircraft surface de-icing during severe winter weather is a crucial step in ensuring flight punctuality and safety. Currently, airports typically employ manually driven de-icing trucks with robotic arms operated by humans for de-icing operations. However, manually driven de-icing trucks have the following significant drawbacks in practical applications: First, de-icing operations typically occur in adverse weather conditions with extremely poor visibility, such as rain, snow, heavy fog, and nighttime. Human piloting relies heavily on the driver's vision and experience, and in low-visibility environments, they are highly susceptible to blind spots, increasing the safety risk of collisions with expensive aircraft or other apron equipment. Second, during extreme weather events such as widespread snowfall, the demand for flight de-icing surges. Prolonged, high-intensity work by pilots can easily lead to fatigue, resulting in decreased operational precision and slower reaction times, further increasing safety risks. Furthermore, it is difficult for human pilots to maintain a high degree of consistency between their approach trajectory and final stopping position when near the target aircraft. This not only affects the efficiency of subsequent robotic de-icing operations and the uniformity of de-icing fluid coverage but also maintains high training and operating costs for personnel.

[0003] To overcome the aforementioned shortcomings of manual driving, the introduction of autonomous de-icing vehicles has become an industry trend. However, compared to autonomous driving on ordinary public roads, achieving autonomous driving of de-icing vehicles in the unique closed environment of airports faces higher technical barriers and challenges: First, airport aprons have extremely high security management levels and special traffic rules. In a real airport environment, service vehicles (even autonomous vehicles) are absolutely not allowed to weave in and out of the apron and must strictly follow the pre-defined ground service roads. This means that conventional free-space-based global path planning algorithms cannot be directly applied.

[0004] Second, airport environments are strictly regulated, limiting the "freedom" of vehicle movement. For example, when an autonomous vehicle encounters a dynamic obstacle on a fixed route, it is typically not allowed to overtake or change lanes at will. The system must be able to make minor adjustments to navigate around obstacles within the strict constraints of the current lane boundaries, or make decisions to slow down or stop and wait.

[0005] Third, the apron environment is highly dynamic and complex. During operation, autonomous de-icing vehicles need to use sensors to accurately identify and avoid various special apron-specific equipment and personnel, such as baggage carts and ground staff, which places extremely high demands on perception and local obstacle avoidance algorithms.

[0006] Fourth, the fault tolerance for terminal docking operations is zero. After arriving at the target aircraft stand, the de-icing truck must eventually leave the fixed service lane and enter the "semi-open area" surrounding the stand. During this terminal docking phase on the apron, the vehicle must precisely generate a "virtual safety envelope" around the aircraft and perform multi-degree-of-freedom fine-tuning at extremely low speeds. Because the target is an extremely expensive and fragile aircraft, any minor trajectory planning error or control delay could lead to a catastrophic collision.

[0007] Therefore, how to combine the airport's inherent road network rules and stringent docking safety requirements to design an intelligent chassis path planning scheme for autonomous de-icing vehicles that can both comply with regulations and follow tracks, and accurately and safely dock for operations, is a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0008] The following provides a brief overview of one or more aspects to offer a basic understanding of them. This overview is not an exhaustive summary of all conceived aspects, nor is it intended to identify key or decisive elements of all aspects, nor to define the scope of any or all aspects. Its sole purpose is to present some concepts of one or more aspects in a simplified form to prepare for the more detailed descriptions that follow.

[0009] The purpose of this invention is to address the above-mentioned needs by providing an intelligent path planning method and system for de-icing truck chassis, enabling automatic driving of the de-icing truck chassis throughout the entire de-icing process.

[0010] The technical solution of this invention is as follows: This invention discloses an intelligent path planning method for a de-icing truck chassis, the method including a global topology path planning stage, a lane-in-lane driving control stage, and an apron endpoint docking stage, wherein, The global topology path planning phase includes: Step 1: Load the high-precision vector map of the airport; Step 2: Convert the high-precision vector map into a global road network topology map; and Step 3: Perform path search based on the global road network topology map; The lane-keeping control phase includes: Step 4: Acquire sensor data and identify dynamic obstacles; Step 5: Conduct local tracking and safety assessment; and Step 6: Perform dynamic obstacle avoidance within the lane; The apron terminal docking phase includes: Step 7: When the de-icing truck reaches the intersection of the target location, it leaves the fixed service lane and enters the semi-open area around the location, switching to the local cost map. Step 8: Generate a virtual safety envelope based on the aircraft type; and Step 9: Use a local path planning algorithm, combined with a virtual safety envelope, to enable the de-icing vehicle to make multi-degree-of-freedom fine adjustments to get closer to the vehicle body.

[0011] According to an embodiment of the intelligent path planning method for the chassis of the de-icing vehicle of the present invention, step 2 further includes: Step 2-1: Extract semantic information strongly related to vehicle movement from the high-precision vector map; Step 2-2: Generate nodes in the global road network topology graph based on semantic information; Steps 2-3: Generate edges in the global road network topology graph based on semantic information; Steps 2-4: Calculate and assign edge weights in the global road network topology graph based on semantic information; Steps 2-5: Construct the data structure of the global road network topology graph using the extracted nodes, edges, and corresponding edge weights.

[0012] According to an embodiment of the intelligent path planning method for the chassis of the de-icing vehicle of the present invention, step 3 further includes: Step 3-1: Starting from the current location of the de-icing truck and ending at the target location, map the location to a node in the global road network topology map; Step 3-2: Run the path search algorithm on the global road network topology map to find the shortest or most time-optimal road segment sequence; Step 3-3: Extract the lane centerline of the road segment sequence found in Step 3-2, and use it as the absolute reference baseline for the subsequent driving of the de-icing vehicle.

[0013] According to an embodiment of the intelligent path planning method for the de-icing truck chassis of the present invention, in step 5, during the lane keeping process, a pure tracking or linear quadratic regulator algorithm is used to calculate the steering angle of the vehicle chassis to ensure that the de-icing truck travels on the center line of the lane.

[0014] According to an embodiment of the intelligent path planning method for the de-icing vehicle chassis of the present invention, in step 6, if there is a dynamic obstacle in the lane ahead, the vehicle is given priority to slow down or stop and wait; if it is determined that the obstacle is stationary and must be detoured, the local path planning algorithm is invoked to make fine adjustments to the detour under the strict constraints of the current lane boundary.

[0015] According to an embodiment of the intelligent path planning method for the chassis of the de-icing vehicle of the present invention, step 8 further includes: Step 8-1: Call the prior model of the aircraft type; Step 8-2: Perform aircraft pose alignment based on real-time point cloud data to obtain the aircraft's precise real-time position and orientation on the tarmac; Step 8-3: Perform geometric dilation calculations on the aligned 3D outer contour of the aircraft to generate a 3D virtual safety envelope; Step 8-4: Project the 3D virtual security envelope onto the local cost map in a reduced dimension, and generate the final 2D virtual security envelope on the local cost map.

[0016] According to an embodiment of the intelligent path planning method for the chassis of the de-icing vehicle of the present invention, in step 9, the local path planning algorithm includes the dynamic window method or the model predictive control method, which sets the cost of all grid areas within the virtual safety envelope to infinity, so that when local path sampling is performed, any simulated trajectory that touches or crosses the virtual safety envelope will be directly eliminated.

[0017] This invention also explains an intelligent path planning system for a de-icing truck chassis. The system includes a global topology path planning module, a lane-in-lane driving control module, and an apron endpoint docking module, wherein... The global topology path planning module includes: The map loading unit loads a high-precision vector map of the airport. The topology transformation unit converts a high-precision vector map into a global road network topology map; and The global path search unit performs path search based on the global road network topology map; The lane-keeping control module includes: The obstacle recognition unit acquires sensor data and identifies dynamic obstacles; Local tracking unit performs local tracking and safety assessment; and The obstacle avoidance unit performs dynamic obstacle avoidance within the lane. The apron terminal docking module includes: The map switching unit, when the de-icing truck reaches the intersection of the target location, leaves the fixed service lane and enters the semi-open area around the location, switches to the local cost map. A virtual safety envelope generation unit generates a virtual safety envelope based on the aircraft type; and The local path planning unit uses a local path planning algorithm combined with a virtual safety envelope to enable the de-icing vehicle to make multi-degree-of-freedom fine adjustments to get closer to the vehicle body.

[0018] The present invention also discloses a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the intelligent path planning method for the de-icing truck chassis as described above.

[0019] The present invention also discloses a computer program product, characterized in that, when the program is executed by a processor, it implements the steps of the intelligent path planning method for the de-icing truck chassis as described above.

[0020] Compared with the prior art, the present invention has the following advantages: 1. Achieve all-weather operation capability and significantly reduce safety hazards caused by personnel fatigue. Addressing the problem of extremely poor visibility and susceptibility to blind spots in adverse weather conditions such as rain, snow, nighttime fog, etc., this application completely eliminates the limitations imposed on the human eye by introducing a multi-sensor real-time perception system.

[0021] 2. Perfectly aligns with stringent airport traffic regulations, ensuring compliant apron driving. Addressing the challenge of applying conventional free-space autonomous driving algorithms directly to apron scenarios, this solution innovatively employs topology map search on a restricted road network. By loading a high-precision vector map containing lane markings and speed limits, the system can search for the optimal lane sequence on the road network, ensuring vehicles strictly adhere to pre-defined ground service lanes and never arbitrarily weave across the apron. Furthermore, in terms of dynamic obstacle avoidance strategies within the road, this solution prioritizes deceleration or stopping, perfectly adapting to the strict discipline and prohibition of unauthorized lane changes and overtaking within airports.

[0022] 3. Zero-tolerance collision avoidance for precise docking, ensuring absolute aircraft safety. Addressing the high challenge of zero tolerance for errors during the apron terminal docking phase, this solution employs a segmented planning logic. Upon reaching the target apron entrance, the system automatically switches to terminal docking mode and reverts to the local cost map. By scanning the target aircraft to generate a dedicated virtual safety envelope, and utilizing the DWA algorithm for multi-degree-of-freedom fine-tuning at extremely low speeds, it finds the final few meters of trajectory to avoid ground support equipment and stay close to the aircraft fuselage. This effectively overcomes the difficulty of precisely controlling distance during blind-spot docking by manual pilots, achieving absolute safety protection for expensive aircraft.

[0023] 4. Improved consistency of operational trajectories, enhanced de-icing efficiency, and reduced operating costs. Addressing the drawback of inconsistent driving trajectories and final parking positions for manual pilots approaching target aircraft, this application's automated path planning accurately extracts the lane centerline as the driving trajectory and calculates the optimal de-icing point upon docking. This highly consistent parking posture provides the ideal starting space for subsequent switching to robotic arm operation mode, significantly improving the robotic arm's de-icing efficiency and de-icing fluid coverage uniformity, while substantially reducing reliance on skilled pilots and high training costs. Attached Figure Description

[0024] The above-described features and advantages of the present invention will be better understood after reading the following detailed description of embodiments of the present disclosure in conjunction with the accompanying drawings. In the drawings, components are not necessarily drawn to scale, and components having similar related characteristics or features may have the same or similar reference numerals.

[0025] Figure 1 A flowchart of an embodiment of the intelligent path planning method for the de-icing vehicle chassis of the present invention is shown.

[0026] Figure 2 It shows Figure 1 The flowchart for step 2 in the method shown.

[0027] Figure 3 It shows Figure 1 The flowchart for step 3 in the method shown.

[0028] Figure 4 It shows Figure 1 The flowchart for step 8 in the method shown.

[0029] Figure 5 A schematic diagram of an embodiment of the intelligent path planning system for de-icing truck chassis of the present invention is shown. Detailed Implementation

[0030] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. It should be noted that the aspects described below with reference to the accompanying drawings and specific embodiments are merely exemplary and should not be construed as limiting the scope of protection of the present invention in any way.

[0031] Figure 1 The flowchart of an embodiment of the intelligent path planning method for the chassis of a de-icing vehicle according to the present invention is shown. Please refer to [link / reference]. Figure 1 The implementation steps of the method in this embodiment are described in detail below.

[0032] The method consists of two phases: global topology path planning, lane-based driving control, and apron terminal docking.

[0033] The global topology path planning phase includes the following processing steps.

[0034] Step 1: Load the high-precision vector map of the airport.

[0035] High-precision vector maps contain a large amount of geometric and semantic information (such as the specific coordinates of lane lines, lane width, and stop line positions), and are usually stored in specific formats (such as OpenDRIVE, Lanelet2, Apollo HD Map, or NDS).

[0036] Step 2: Convert the high-precision vector map into a global road network topology map.

[0037] Combination Figure 2 As shown, step 2 further includes the following processing steps: Step 2-1: Extract semantic information strongly related to vehicle driving from the high-precision vector map. This semantic information includes: lane centerlines, which represent the baseline trajectory of vehicles traveling within the lane; lane connectivity, which clarifies which lanes are connected at intersections (preceding / following relationships); lane boundary attributes, which distinguish between solid lines (cannot be crossed) and dashed lines (reversible lanes); and traffic rule constraints, which include speed limits, traffic light locations, and turning restrictions (such as prohibition of left turns).

[0038] Step 2-2: Generate nodes in the global road network topology map based on semantic information.

[0039] In a topology graph, nodes typically represent "decision points where vehicles may change their driving state" or "discrete waypoints in space." Common node generation methods include: intersections / connection points, i.e., the start and end points of lanes, especially points entering and leaving intersections; lane segmentation points, where, for the sake of planning granularity, a series of intermediate nodes are usually sampled along the lane centerline at fixed intervals (e.g., every 10 or 50 meters); and special semantic points, such as stop line positions and the leading edge of pedestrian crossings.

[0040] Steps 2-3: Generate edges in the global road network topology graph based on semantic information.

[0041] An edge represents a directional drivable path from one node to another. Based on the connectivity of the high-precision map, the following three types of edges are typically generated: Lane Edges, which connect two adjacent nodes within the same lane, with the direction along the traffic flow; Lane Change Edges, which are lateral edges generated between parallel nodes of adjacent lanes when the boundary between them is a dashed line (allowing lane changes); and Junction / Turning Edges, which connect entry lane nodes to legal exit lane nodes within an intersection (such as left-turn trajectories, straight-ahead trajectories, and right-turn trajectories).

[0042] Steps 2-4: Calculate and assign edge weights in the global road network topology graph based on semantic information.

[0043] Each edge is assigned a weight (Cost). The calculation of the weight usually takes into account several factors: the base distance cost, i.e., the physical length of the edge; the time cost (speed limit), i.e., the edge length divided by the maximum speed limit of the road segment. The higher the speed limit, the lower the time cost, and the more the algorithm tends to choose the segment; the turning penalty cost, to avoid frequent steering wheel turns, the cost of straight edges is usually lower than that of left or right turns, and the cost of U-turns is the highest; the lane change penalty cost, which adds an additional fixed cost to lane change edges to prevent the algorithm from planning a path that "draws a dragon" between two lanes (frequent lane changes).

[0044] Steps 2-5: Construct the extracted nodes, edges, and corresponding edge weights into a data structure for the global road network topology graph, typically an adjacency list or adjacency matrix.

[0045] Step 3: Perform path search based on the global road network topology map.

[0046] Combination Figure 3 As shown, step 3 further includes the following processing: Step 3-1: Starting from the current location of the de-icing truck and ending at the target location, map the location to a node in the global road network topology map.

[0047] Step 3-2: Run the path search algorithm on the global road network topology map to find the shortest or most time-optimal road segment sequence, rather than a discrete coordinate point sequence.

[0048] Path search algorithms include, for example, Dijkstra's algorithm or A. algorithm.

[0049] Step 3-3: Extract the lane centerline of the road segment sequence found in Step 3-2, and use it as the absolute reference baseline for the subsequent driving of the de-icing vehicle.

[0050] The lane-keeping control phase includes the following processing steps.

[0051] Step 4: Acquire sensor data and identify dynamic obstacles.

[0052] Sensors include, for example, lidar and cameras on de-icing vehicles, while dynamic obstacles include, for example, luggage carts and ground staff.

[0053] Step 5: Conduct local tracking and safety assessment.

[0054] The first step is lane keeping, which typically uses Pure Pursuit or LQR (Linear Quadratic Regulator) algorithms to calculate the vehicle chassis steering angle and ensure that the de-icing truck stays strictly on the center line of the lane.

[0055] Secondly, longitudinal speed control dynamically adjusts the linear speed of the de-icing truck based on road network speed limits and the distance to the vehicle in front.

[0056] Step 6: Dynamic obstacle avoidance within the lane.

[0057] If there is a dynamic obstacle in the lane ahead, the first priority is to slow down or stop and wait (airports have strict rules and usually do not allow overtaking by changing lanes at will). If it is determined that the obstacle is stationary and must be detoured, then the DWA (Dynamic Window Approach) or MPC (Model Predictive Control) algorithm is called to make fine adjustments to the detour under the strict constraints of the current lane boundary.

[0058] The apron terminal docking phase includes the following processing steps.

[0059] Step 7: When the de-icing truck reaches the intersection of the target location, it leaves the fixed service lane and enters the semi-open area around the location, switching to the local cost map.

[0060] A local cost map is a small-scale dynamic grid map centered on the vehicle and moving with it. It doesn't care about the overall movement of the vehicle; it only projects obstacle information scanned in real time by the vehicle's sensors (LiDAR, cameras) onto this small map and assigns different "costs" to the areas around the obstacles (the closer to the obstacle, the higher the cost and the greater the probability of a collision).

[0061] Step 8: Generate a virtual security envelope based on the aircraft type.

[0062] In airport de-icing scenarios, aircraft are extremely expensive and their surfaces are covered with fragile sensors (such as pitot tubes and hydrostatic orifices). The de-icing truck has zero tolerance for error when approaching an aircraft. Therefore, the system cannot simply treat the aircraft as a simple "point" or "block" to avoid; instead, it must extend an absolutely no-entry boundary distance (e.g., 1.5 meters or 2 meters) around the aircraft's actual physical contours. This boundary line, generated in the algorithmic map, is the virtual safety envelope.

[0063] Combination Figure 4 As shown, step 8 further includes the following processing steps.

[0064] Step 8-1: Call the prior model of the aircraft type.

[0065] By accessing the airport's A-CDM (Airport Collaborative Decision Making System) to obtain flight information, or by identifying the aircraft type (such as Boeing 737 or Airbus A320) through vehicle-mounted cameras / radar, the high-precision 3D geometric model (CAD model or mesh model) corresponding to the aircraft type can be retrieved from the local or cloud-based aircraft type database.

[0066] Step 8-2: Perform aircraft pose alignment based on real-time point cloud data to obtain the aircraft's precise real-time position and orientation on the tarmac.

[0067] The vehicle-mounted LiDAR on the de-icing truck scans the aircraft ahead to acquire real-time local 3D point cloud data. Using registration algorithms such as Iterative Closest Point, the incomplete contours scanned by the radar are perfectly overlapped and matched with the complete 3D model loaded in step 8-1 to obtain extremely accurate degrees of freedom pose of the aircraft in the local coordinate system of the de-icing truck.

[0068] Step 8-3: Perform geometric expansion calculations on the aligned 3D outer contour of the aircraft to generate a 3D virtual safety envelope.

[0069] First, a safety threshold distance is set. Then, using the Minkowski Sum or region dilation algorithm in geometry, each point on the outer surface of the aircraft is pushed outward along its normal vector by this safety threshold distance. This creates a larger "fat contour," or virtual safety envelope, on top of the original contour that fits the fuselage.

[0070] Step 8-4: Project the 3D virtual security envelope onto the local cost map in a reduced dimension, and generate the final 2D virtual security envelope on the local cost map.

[0071] Step 9: Use a local path planning algorithm, combined with a 2D virtual safety envelope, to enable the de-icing truck to make multi-degree-of-freedom fine adjustments at low speeds to approach but not touch the vehicle body.

[0072] Local path planning algorithms include, for example, the Dynamic Window Approach (DWA) or Model Predictive Control (MPC). In the local cost map, the "cost" of all grid areas within this virtual safety envelope is set to infinity. When sampling local paths, any simulated trajectory that touches or crosses this line is directly discarded, thus physically eliminating the possibility of the de-icing truck chassis colliding with the aircraft.

[0073] Figure 5 The principle of an embodiment of the intelligent path planning system for the de-icing truck chassis of the present invention is illustrated. Please refer to [link / reference]. Figure 5 The system in this embodiment includes: a global topology path planning module, a lane driving control module, and an apron endpoint docking module.

[0074] The global topology path planning module includes: map loading unit, topology transformation unit, and global path search unit.

[0075] The map loading unit loads a high-precision vector map of the airport.

[0076] High-precision vector maps contain a large amount of geometric and semantic information (such as the specific coordinates of lane lines, lane width, and stop line positions), and are usually stored in specific formats (such as OpenDRIVE, Lanelet2, Apollo HD Map, or NDS).

[0077] The topology transformation unit converts a high-precision vector map into a global road network topology map.

[0078] The topology conversion unit is further configured to perform the following processing steps.

[0079] Semantic information strongly related to vehicle movement is extracted from high-precision vector maps. This semantic information includes: lane centerlines, which represent the baseline trajectory of vehicles traveling within the lane; lane connectivity, which clarifies which lanes are connected at intersections (preceding / following relationships); lane boundary attributes, which distinguish between solid lines (cannot be crossed) and dashed lines (reversible lanes); and traffic rule constraints, which include speed limits, traffic light locations, and turning restrictions (such as prohibition of left turns).

[0080] Nodes are generated in the global road network topology map based on semantic information.

[0081] In a topology graph, nodes typically represent "decision points where vehicles may change their driving state" or "discrete waypoints in space." Common node generation methods include: intersections / connection points, i.e., the start and end points of lanes, especially points entering and leaving intersections; lane segmentation points, where, for the sake of planning granularity, a series of intermediate nodes are usually sampled along the lane centerline at fixed intervals (e.g., every 10 or 50 meters); and special semantic points, such as stop line positions and the leading edge of pedestrian crossings.

[0082] Edges are generated in the global road network topology graph based on semantic information.

[0083] An edge represents a directional drivable path from one node to another. Based on the connectivity of the high-precision map, the following three types of edges are typically generated: Lane Edges, which connect two adjacent nodes within the same lane, with the direction along the traffic flow; Lane Change Edges, which are lateral edges generated between parallel nodes of adjacent lanes when the boundary between them is a dashed line (allowing lane changes); and Junction / Turning Edges, which connect entry lane nodes to legal exit lane nodes within an intersection (such as left-turn trajectories, straight-ahead trajectories, and right-turn trajectories).

[0084] Edge weights in the global road network topology graph are calculated and assigned based on semantic information.

[0085] Each edge is assigned a weight (Cost). The calculation of the weight usually takes into account several factors: the base distance cost, i.e., the physical length of the edge; the time cost (speed limit), i.e., the edge length divided by the maximum speed limit of the road segment. The higher the speed limit, the lower the time cost, and the more the algorithm tends to choose the segment; the turning penalty cost, to avoid frequent steering wheel turns, the cost of straight edges is usually lower than that of left or right turns, and the cost of U-turns is the highest; the lane change penalty cost, which adds an additional fixed cost to lane change edges to prevent the algorithm from planning a path that "draws a dragon" between two lanes (frequent lane changes).

[0086] The extracted nodes, edges, and corresponding edge weights are used to construct a data structure for the global road network topology graph, typically an adjacency list or adjacency matrix.

[0087] The global path search unit performs path search based on the global road network topology map.

[0088] The global path search unit is further configured to perform the following processing steps: Starting from the current location of the de-icing truck and ending at the target location, the data is mapped to a node in the global road network topology map.

[0089] The path search algorithm runs on the global road network topology map to find the shortest or most time-optimal road sequence, rather than a discrete sequence of coordinate points.

[0090] Path search algorithms include, for example, Dijkstra's algorithm or A. algorithm.

[0091] Extract the lane centerlines from the road segment sequence found in the previous step, and use them as the absolute reference baseline for the subsequent driving of the de-icing truck.

[0092] The lane-keeping control module includes: an obstacle recognition unit, a local tracking unit, and an obstacle avoidance unit.

[0093] The obstacle recognition unit acquires sensor data and identifies dynamic obstacles. Sensors include, for example, lidar and cameras on a de-icing vehicle, while dynamic obstacles include, for example, luggage carts and ground staff.

[0094] The local tracking unit performs local tracking and safety assessments. For lane keeping, Pure Pursuit or LQR (Linear Quadratic Regulator) algorithms are typically used to calculate the vehicle's steering angle, ensuring the de-icing truck stays strictly on the lane centerline. For longitudinal speed control, the de-icing truck's linear speed is dynamically adjusted based on road network speed limits and the distance to the vehicle ahead.

[0095] The obstacle avoidance unit performs dynamic obstacle avoidance within the lane. If there is a dynamic obstacle in the lane ahead, it prioritizes slowing down or stopping to wait (airport regulations are strict, and overtaking is usually not allowed). If it determines that the obstacle is stationary and detour is necessary, it then calls the DWA (Dynamic Window Approach) or MPC (Model Predictive Control) algorithm to fine-tune the detour under the strict constraints of the current lane boundary.

[0096] The apron endpoint docking module includes: a map switching unit, a virtual safety envelope generation unit, and a local path planning unit.

[0097] The map switching unit, when the de-icing truck arrives at the intersection bordering the target location, leaves the fixed service lane and enters the semi-open area around the location, switching to the local cost map.

[0098] A local cost map is a small-scale dynamic grid map centered on the vehicle and moving with it. It doesn't care about the overall movement of the vehicle; it only projects obstacle information scanned in real time by the vehicle's sensors (LiDAR, cameras) onto this small map and assigns different "costs" to the areas around the obstacles (the closer to the obstacle, the higher the cost and the greater the probability of a collision).

[0099] The virtual safety envelope generation unit generates a virtual safety envelope based on the aircraft model.

[0100] In airport de-icing scenarios, aircraft are extremely expensive and their surfaces are covered with fragile sensors (such as pitot tubes and hydrostatic orifices). The de-icing truck has zero tolerance for error when approaching an aircraft. Therefore, the system cannot simply treat the aircraft as a simple "point" or "block" to avoid; instead, it must extend an absolutely no-entry boundary distance (e.g., 1.5 meters or 2 meters) around the aircraft's actual physical contours. This boundary line, generated in the algorithmic map, is the virtual safety envelope.

[0101] The virtual security envelope generation unit is further configured to perform the following processing steps.

[0102] The system retrieves the aircraft type prior model. It obtains flight information by accessing the airport's A-CDM (Airport Collaborative Decision Making System) or identifies the aircraft type (such as Boeing 737 or Airbus A320) through vehicle-mounted cameras / radar. Then, it retrieves the corresponding high-precision 3D geometric model (CAD model or mesh model) from the local or cloud-based aircraft type database.

[0103] The aircraft's pose is aligned using real-time point clouds to obtain its precise real-time position and orientation on the tarmac. A vehicle-mounted LiDAR scanner on the de-icing truck scans the aircraft ahead, acquiring real-time local 3D point cloud data. Registration algorithms such as Iterative Closest Point (ITP) are used to perfectly overlap and match the incomplete contours scanned by the radar with the complete 3D model loaded in step 8-1, resulting in an extremely accurate degree-of-freedom pose of the aircraft in the local coordinate system of the de-icing truck.

[0104] Geometric dilation calculations are performed on the aligned 3D aircraft outline to generate a 3D virtual safety envelope. First, a safety threshold distance is set. Then, using the Minkowski Sum or region dilation algorithm from geometry, each point on the aircraft's outer surface is shifted outward along its normal vector by this safety threshold distance. This creates a larger "fat outline," or virtual safety envelope, on top of the original outline that fits the fuselage.

[0105] The 3D virtual security envelope is reduced in dimension and projected onto the local cost map, generating the final 2D virtual security envelope on the local cost map.

[0106] The local path planning unit uses a local path planning algorithm combined with a virtual safety envelope to enable the de-icing vehicle to make multi-degree-of-freedom fine adjustments to get closer to the vehicle body.

[0107] Local path planning algorithms include, for example, the Dynamic Window Approach (DWA) or Model Predictive Control (MPC). In the local cost map, the "cost" of all grid areas within this virtual safety envelope is set to infinity. When sampling local paths, any simulated trajectory that touches or crosses this line is directly discarded, thus physically eliminating the possibility of the de-icing truck chassis colliding with the aircraft.

[0108] Furthermore, this invention also discloses a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the following... Figure 1 The steps of the intelligent path planning method embodiment for de-icing truck chassis shown are illustrated.

[0109] Furthermore, this invention also discloses a computer program product, which, when executed by a processor, implements the following: Figure 1 The steps of the intelligent path planning method embodiment for de-icing truck chassis shown are illustrated.

[0110] Although the methods described above are illustrated and depicted as a series of actions for the sake of simplicity, it should be understood and appreciated that these methods are not limited by the order of the actions, as some actions may occur in a different order and / or concurrently with other actions from the illustrations and descriptions herein or not illustrated and described herein but which may be understood by those skilled in the art, according to one or more embodiments.

[0111] Those skilled in the art will further appreciate that the various illustrative logic blocks, modules, circuits, and algorithm steps described in conjunction with the embodiments disclosed herein can be implemented as electronic hardware, computer software, or a combination of both. To clearly illustrate this interchangeability between hardware and software, the various illustrative components, blocks, modules, circuits, and steps are described above in a generalized manner in terms of their functionality. Whether such functionality is implemented as hardware or software depends on the specific application and the design constraints imposed on the overall system. Those skilled in the art may implement the described functionality in different ways for each specific application, but such implementation decisions should not be construed as departing from the scope of the invention.

[0112] The various illustrative logic blocks, modules, and circuits described in conjunction with the embodiments disclosed herein can be implemented or performed using a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. The general-purpose processor may be a microprocessor, but in alternatives, it may be any conventional processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors cooperating with a DSP core, or any other such configuration.

[0113] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of both. The software module may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to a processor such that the processor can read and write information to / from the storage medium. In an alternative, the storage medium may be integrated into the processor. The processor and storage medium may reside in an ASIC. The ASIC may reside in a user terminal. In an alternative, the processor and storage medium may reside as discrete components in the user terminal.

[0114] In one or more exemplary embodiments, the described functionality may be implemented in hardware, software, firmware, or any combination thereof. If implemented in software as a computer program product, the functionality may be stored or transmitted as one or more instructions or code on or through a computer-readable medium. A computer-readable medium includes both computer storage media and communication media, encompassing any medium that facilitates the transfer of a computer program from one location to another. A storage medium may be any available medium accessible to a computer. By way of example and not limitation, such a computer-readable medium may include RAM, ROM, EEPROM, CD-ROM or other optical disc storage, disk storage or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and is accessible to a computer. Any connection is also legitimately referred to as a computer-readable medium. For example, if the software is transmitted from a website, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave, then the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of a medium. As used in this article, disk and disc include compact discs (CDs), laser discs, optical discs, digital multi-purpose discs (DVDs), floppy disks, and Blu-ray discs. Disks typically reproduce data magnetically, while discs reproduce data optically using lasers. Combinations of these should also be included within the scope of computer-readable media.

[0115] The prior description of this disclosure is provided to enable any person skilled in the art to make or use this disclosure. Various modifications to this disclosure will be apparent to those skilled in the art, and the general principles defined herein may be applied to other variations without departing from the spirit or scope of this disclosure. Therefore, this disclosure is not intended to be limited to the examples and designs described herein, but should be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for intelligent path planning of a de-icing truck chassis, characterized in that, The method includes a global topology path planning phase, a lane-based driving control phase, and an apron endpoint docking phase, among which... The global topology path planning phase includes: Step 1: Load the high-precision vector map of the airport; Step 2: Convert the high-precision vector map into a global road network topology map; and Step 3: Perform path search based on the global road network topology map; The lane-keeping control phase includes: Step 4: Acquire sensor data and identify dynamic obstacles; Step 5: Conduct local tracking and safety assessment; and Step 6: Perform dynamic obstacle avoidance within the lane; The apron terminal docking phase includes: Step 7: When the de-icing truck reaches the intersection of the target location, it leaves the fixed service lane and enters the semi-open area around the location, switching to the local cost map. Step 8: Generate a virtual safety envelope based on the aircraft type; and Step 9: Use a local path planning algorithm, combined with a virtual safety envelope, to enable the de-icing vehicle to make multi-degree-of-freedom fine adjustments to get closer to the vehicle body.

2. The intelligent path planning method for de-icing truck chassis according to claim 1, characterized in that, Step 2 further includes: Step 2-1: Extract semantic information strongly related to vehicle movement from the high-precision vector map; Step 2-2: Generate nodes in the global road network topology graph based on semantic information; Steps 2-3: Generate edges in the global road network topology graph based on semantic information; Steps 2-4: Calculate and assign edge weights in the global road network topology graph based on semantic information; Steps 2-5: Construct the data structure of the global road network topology graph using the extracted nodes, edges, and corresponding edge weights.

3. The intelligent path planning method for de-icing truck chassis according to claim 1, characterized in that, Step 3 further includes: Step 3-1: Starting from the current location of the de-icing truck and ending at the target location, map the location to a node in the global road network topology map; Step 3-2: Run the path search algorithm on the global road network topology map to find the shortest or most time-optimal road segment sequence; Step 3-3: Extract the lane centerline of the road segment sequence found in Step 3-2, and use it as the absolute reference baseline for the subsequent driving of the de-icing vehicle.

4. The intelligent path planning method for de-icing truck chassis according to claim 1, characterized in that, In step 5, during lane keeping, a pure tracking or linear quadratic regulator algorithm is used to calculate the steering angle of the vehicle chassis to ensure that the de-icing vehicle stays on the center line of the lane.

5. The intelligent path planning method for de-icing truck chassis according to claim 1, characterized in that, In step 6, if there is a dynamic obstacle in the lane ahead, the priority is to slow down or stop and wait; if it is determined that the obstacle is stationary and detour is necessary, the local path planning algorithm is invoked to make fine adjustments to the detour under the strict constraints of the current lane boundary.

6. The intelligent path planning method for de-icing truck chassis according to claim 1, characterized in that, Step 8 further includes: Step 8-1: Call the prior model of the aircraft type; Step 8-2: Perform aircraft pose alignment based on real-time point cloud data to obtain the aircraft's precise real-time position and orientation on the tarmac; Step 8-3: Perform geometric dilation calculations on the aligned 3D outer contour of the aircraft to generate a 3D virtual safety envelope; Step 8-4: Project the 3D virtual security envelope onto the local cost map in a reduced dimension, and generate the final 2D virtual security envelope on the local cost map.

7. The intelligent path planning method for de-icing truck chassis according to claim 1, characterized in that, In step 9, the local path planning algorithm includes the dynamic window method or the model predictive control method, which sets the cost of all grid areas within the virtual safety envelope to infinity, so that when local path sampling is performed, any simulated trajectory that touches or crosses the virtual safety envelope will be directly eliminated.

8. An intelligent path planning system for a de-icing truck chassis, characterized in that, The system includes a global topology path planning module, a lane-based driving control module, and an apron endpoint docking module. The global topology path planning module includes: The map loading unit loads a high-precision vector map of the airport. The topology transformation unit converts a high-precision vector map into a global road network topology map; and The global path search unit performs path search based on the global road network topology map; The lane-keeping control module includes: The obstacle recognition unit acquires sensor data and identifies dynamic obstacles; Local tracking units perform local tracking and safety assessments; and The obstacle avoidance unit performs dynamic obstacle avoidance within the lane. The apron terminal docking module includes: The map switching unit, when the de-icing truck reaches the intersection of the target location, leaves the fixed service lane and enters the semi-open area around the location, switches to the local cost map. A virtual safety envelope generation unit generates a virtual safety envelope based on the aircraft type; and The local path planning unit uses a local path planning algorithm combined with a virtual safety envelope to enable the de-icing vehicle to make multi-degree-of-freedom fine adjustments to get closer to the vehicle body.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the intelligent path planning method for the de-icing vehicle chassis as described in any one of claims 1-7.

10. A computer program product, characterized in that, When the program is executed by the processor, it implements the steps of the intelligent path planning method for the de-icing truck chassis as described in any one of claims 1-7.

Images