A collision computation acceleration method of a parameterized geometry combined collision algorithm

Abstract

This invention discloses a collision calculation acceleration method combining parametric geometry and collision algorithms, belonging to the field of physical simulation technology. The method includes: traversing a 3D scene to classify and label geometric objects as parametric entities or discrete mesh objects; constructing a hybrid bounding volume hierarchical structure tree, with leaf nodes storing algebraic equations or vertex data respectively; performing generalized collision detection to screen potential collision pairs; calculating a collision urgency index based on overlapping volume and complexity, and prioritizing collision pairs; using analytical geometry for parametric entities and primitive intersection testing for discrete meshes, and dynamically allocating computational resources according to priority; finally, using pose hash caching and a warm-start strategy to accelerate convergence. This invention effectively solves the problems of high computational load and inefficient resource allocation in traditional discretized collision detection by combining hybrid geometric representation and dynamic scheduling mechanisms, significantly improving the real-time performance of physical simulation.

Description

Technical Field

[0001] This invention relates to the field of physical simulation technology, and in particular to a method for accelerating collision calculation by combining parametric geometry with a collision algorithm. Background Technology

[0002] In applications such as virtual reality, 3D game development, and industrial simulation, collision detection is one of the most core and computationally intensive modules in a physics engine. However, existing mainstream collision detection algorithms typically employ a single discretization logic, meaning that regardless of whether objects in the scene are regular geometric shapes or complex freeform surfaces, they are all discretized into a mesh model composed of numerous triangular facets for calculation. This "one-size-fits-all" approach has significant drawbacks: on the one hand, it discards the inherent simplicity of algebraic features of regular geometric shapes, degenerating a simple analytical geometry problem into a tedious test of massive primitive intersections, leading to an exponential increase in computational load and accuracy limited by mesh resolution; on the other hand, existing generalized and narrow collision stages often lack dynamic scheduling mechanisms based on physical states, typically processing all potential collision pairs in a linear order, failing to differentiate computational priorities based on the proximity or geometric complexity between objects. This results in a significant waste of processor resources on low-probability or non-urgent collision calculations, making it difficult to meet the real-time interactive requirements of high frame rates and low latency in large-scale complex scenes. Summary of the Invention

[0003] To address the aforementioned problems in the existing technology, the present invention aims to provide a method for accelerating collision calculation by combining parametric geometry with a collision algorithm, the method comprising the following steps: Step S1: Obtain the 3D scene data to be collision detected, and traverse all geometric objects in the 3D scene data. Classify and label the geometric objects as parametric solid objects or discrete mesh objects according to their geometric features.

[0004] Step S2: Construct a hybrid bounding volume hierarchical structure tree. The leaf nodes of the hybrid bounding volume hierarchical structure tree store the algebraic equation parameters of the parameterized entity object and the vertex data of the discrete mesh object, respectively.

[0005] Step S3: Traverse the hybrid bounding volume hierarchical structure tree to perform generalized stage collision detection and filter out a list of potential collision pairs.

[0006] Step S4: Prioritize the list of potential collision pairs based on the collision urgency index, and perform precise collision calculations in the narrow stage in sequence. The analytical geometry solution method is used for the parameterized entity object, and the primitive intersection test method is used for the discrete mesh object.

[0007] Step S5: Output the collision detection results and update the physical state of the 3D scene data based on the collision detection results.

[0008] Preferably, in step S1, classifying and labeling the geometric object as a parametric solid object or a discrete mesh object based on its geometric features includes the following sub-steps: Step S101: Detect whether the shape topological features of the current geometric object match the preset basic body library, which includes at least spheres, cuboids, cylinders and cones.

[0009] Step S102: If the matching is successful, the geometric center, semi-axis length and rotation quaternion of the geometric object are extracted as parameters of the algebraic equation, and the geometric object is marked as the parameterized entity object.

[0010] Step S103: If the matching fails, obtain the face index and vertex array of the geometric object, and mark the geometric object as the discrete mesh object.

[0011] Preferably, in step S4, the collision urgency index is used to quantify the calculation priority of each potential collision pair, and its calculation formula is as follows: ,in, This represents the collision urgency index; the higher the index, the higher the calculation priority. This represents the volume of the overlapping region of the axis-aligned bounding boxes of the two objects in the potential collision pair; and These represent the geometric center coordinate vectors of the two objects in the potential collision pair, respectively. Represents the square operation of Euclidean distance; To prevent the denominator from being zero, a preset minimum distance constant is used; and These represent the geometric computational complexity coefficients corresponding to the two objects. If the object is marked as the parameterized entity object, the coefficient takes a preset low value; if the object is marked as the discrete mesh object, the coefficient takes a preset high value. and The preset weighted balance factor; This represents the global normalization constant.

[0012] Preferably, in step S4, when both objects in the potential collision pair are the parameterized entity objects, the precise collision calculation in the narrow phase includes the following sub-steps: Step S401: Read the local coordinate system transformation matrices of the two parameterized entity objects.

[0013] Step S402: Transform the algebraic equation of one of the parameterized entity objects into the local coordinate system of the other parameterized entity object.

[0014] Step S403: Construct the minimum distance analytical equation between the two parameterized entity objects.

[0015] Step S404: Solve for the root of the minimum distance analytical equation using Newton's iteration method. If the root of the minimum distance analytical equation is less than or equal to zero, then a collision is determined to have occurred.

[0016] Preferably, step S2 includes the following sub-steps: Step S201: Construct the hybrid bounding volume hierarchical structure tree in a top-down manner.

[0017] In step S202, during the node splitting process, geometric objects that are spatially adjacent and belong to the same parameterized entity object are preferentially aggregated to the same parent node.

[0018] Step S203: Calculate the bounding box for each node. If the node contains the parameterized entity object, generate a compact bounding box directly based on the algebraic equation parameters to replace vertex-based bounding box calculation.

[0019] Preferably, in step S4, when the potential collision pair includes one parameterized entity object and one discrete mesh object, the precise collision calculation in the narrow stage includes: obtaining a set of local facets in the discrete mesh object that are located within the bounding box of the parameterized entity object; calculating the distance from each triangular facet in the set of local facets to the geometric center of the parameterized entity object; and determining whether the distance is less than a surface-defined threshold of the parameterized entity object to determine the collision penetration depth.

[0020] Preferably, the specific execution logic of step S4 is based on the collision urgency index. Dynamic scheduling includes: establishing high-priority queues and ordinary-priority queues; and setting the collision urgency index... Potential collision pairs exceeding a preset urgency threshold are pushed into the high-priority queue, and the remaining potential collision pairs are pushed into the normal-priority queue. For tasks in the high-priority queue, main processor core resources are allocated to prioritize the execution of the narrow-stage collision precision calculation. For tasks in the normal-priority queue, background threads or idle computing cycles are allocated to perform delayed calculations. When a potential collision pair contains at least one of the parameterized entity objects and the collision urgency index... When the value is extremely high, the contact point of the previous time step is directly used as the initial value for hot start to perform fast convergence calculation.

[0021] Preferably, step S3 includes continuous collision detection logic, which includes the following sub-steps: Step S301: Based on the motion velocity vector of the previous frame, generate a scan bounding volume stretched along the motion path for the parameterized entity object.

[0022] Step S302: Detect the overlap between the scanned bounding volume and the bounding volumes of other objects.

[0023] Step S303: Add the corresponding object combination to the potential collision pair list only when the scan bounding volumes overlap.

[0024] Preferably, in step S4, when using the analytical geometry solution method for the parameterized entity object, a caching mechanism is introduced, including: using the relative pose matrix of the two parameterized entity objects as hash keys; querying the historical collision cache table, and if a corresponding hash key exists, directly reading the distance calculation result of the previous time step as the initial value; and performing a small number of iterative correction calculations based on the initial value to obtain the current collision state.

[0025] Preferably, step S1 further includes a preprocessing operation on the hybrid geometry, including: Identify target geometric objects that are composed of complex discrete meshes but whose overall outline approximates standard geometry; fit the target geometric object with the least squares method to obtain the maximum inscribed parameterized entity and the minimum circumscribed parameterized entity; in step S4, the maximum inscribed parameterized entity and the minimum circumscribed parameterized entity are used first for fast elimination test, and the original discrete mesh data is called only when the test results are unclear.

[0026] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention constructs a hybrid geometric representation and differentiated solution system. By automatically identifying regular geometric objects in the scene and preserving their algebraic equation characteristics, it avoids the computational redundancy and accuracy loss caused by the traditional algorithm's forced discretization of all objects into a high-density grid. The system uses an analytical geometry method based on Newton's iteration to solve parametric entities and primitive intersection testing for discrete grids. Combined with hybrid bounding volume hierarchical structure tree and inscribed / circumscribed fitting preprocessing, it significantly reduces memory usage and greatly improves the collision detection computation efficiency in complex scenes while ensuring high accuracy of physical simulation.

[0027] This invention innovatively introduces a collision urgency index and a dynamic resource scheduling mechanism, which can comprehensively measure the computational priority of each potential collision pair by considering overlapping volume, distance approximation, and geometric complexity. This establishes high-priority and ordinary-priority queues to distribute computational tasks. Combined with a hash cache based on historical poses and a warm-start strategy, the system uses the contact point of the previous frame as an initial value for rapid convergence, realizing the transformation from linear traversal to on-demand computation. This effectively alleviates the computational bottleneck in large-scale scenarios and ensures the real-time performance and stability of physical interactions in highly dynamic environments. Attached Figure Description

[0028] Figure 1 This is an exemplary flowchart of the acceleration method of the present invention.

[0029] Figure 2 This is an exemplary flowchart of the steps for classifying and marking geometric features according to the present invention.

[0030] Figure 3 This is an exemplary flowchart of the collision calculation steps of the present invention.

[0031] Figure 4 An exemplary flowchart illustrating the steps of constructing the structure tree for this invention.

[0032] Figure 5 This is an exemplary flowchart of the continuous collision monitoring logic of the present invention. Detailed Implementation

[0033] The present invention will be further described below with reference to specific embodiments.

[0034] like Figure 1 As shown in this embodiment, a collision calculation acceleration method combining parametric geometry and collision algorithm is provided. The method includes the following steps: Step S1: Acquire the 3D scene data to be collided with, and traverse all geometric objects in the 3D scene data. Based on geometric features, classify and label the geometric objects as parametric solid objects or discrete mesh objects. In practice, the 3D scene data can originate from scene graphs in a game engine or boundary representation data from industrial simulation software. The system first establishes an object attribute table and determines its type by parsing the object's metadata or geometric construction history. For example, regular parts generated by features such as stretching and rotation are more likely to be labeled as parametric solid objects, while freeform surface models generated through 3D scanning or sculpting are labeled as discrete mesh objects.

[0035] like Figure 2 As shown, in step S1 of this embodiment, the geometric objects are classified and labeled as parametric solid objects or discrete mesh objects according to their geometric features, including the following sub-steps: Step S101: Detect whether the shape and topological features of the current geometric object match a preset basic geometry library. The basic geometry library includes at least spheres, cuboids, cylinders, and cones. Specifically, the system extracts the vertex distribution features, curvature histogram, or skeleton structure of the object. For example, if the variance of the distances from all vertices to a certain point is extremely small, it matches as a sphere; if the vertices are distributed on two parallel planes and the side normals are radially distributed, it matches as a cylinder. The basic geometry library can also be expanded to include common geometric shapes such as capsules and tori.

[0036] In step S102, if the matching is successful, the geometric center, semi-axis length, and rotation quaternion of the geometric object are extracted as parameters of the algebraic equation, and the geometric object is marked as a parameterized entity object. At this time, the object no longer needs to store a large number of triangular faces in memory, but only a small number of equation coefficients, thus significantly reducing memory usage. Compared with Euler angles, rotation quaternions can effectively avoid gimbal lock problems.

[0037] In step S103, if the matching fails, the face indices and vertex arrays of the geometric object are obtained, and the geometric object is marked as a discrete mesh object. For such objects, the system will use a half-edge structure or a compressed vertex array for storage, and construct an axis-aligned bounding box or a directed bounding box for it to accelerate subsequent processing.

[0038] Step S1 also includes preprocessing operations on the hybrid geometry, including: The process identifies target geometric objects composed of complex discrete meshes, but whose overall outline approximates standard geometry. The least squares method is used to fit the maximum inscribed parametric entity and the minimum bounding parametric entity of the target geometric object. In step S4, the maximum inscribed parametric entity and the minimum bounding parametric entity are used preferentially for rapid elimination testing; the original discrete mesh data is only called when the test results are unclear. If the minimum bounding entities of two objects do not intersect, then the original objects definitely do not intersect, achieving rapid elimination; if the maximum inscribed entities of two objects intersect, then the original objects definitely intersect, achieving rapid confirmation. Only in the ambiguous area where the bounding entities intersect but the inscribed entities do not, is it necessary to call expensive mesh computing resources.

[0039] Step S2 involves constructing a hybrid bounding volume hierarchical tree structure. The leaf nodes of this structure store the algebraic equation parameters of the parameterized solid objects and the vertex data of the discrete mesh objects, respectively. This tree structure typically employs a bounding volume hierarchy (BVH) or a KD tree. Unlike a traditional BVH, this is a heterogeneous tree: for parameterized leaf nodes, geometric parameters are directly stored; for mesh leaf nodes, triangle indices are stored.

[0040] like Figure 4 As shown, step S2 in this embodiment includes the following sub-steps: Step S201: Construct a hybrid bounding volume hierarchical structure tree in a top-down manner; use the surface area heuristic algorithm SAH to find the optimal splitting plane to minimize the traversal overhead of ray or collision detection.

[0041] In step S202, during node splitting, spatially adjacent geometric objects belonging to the same parametric entity object are preferentially aggregated under the same parent node. Vertex-based bounding boxes tend to be large, while bounding boxes generated based on equations are compact. Compact bounding boxes can significantly reduce the false alarm rate during tree traversal, thereby reducing the number of times leaf nodes are explored.

[0042] Step S203: Calculate the bounding box for each node. If the node contains a parameterized entity object, a compact bounding box is directly generated based on the parameters of the algebraic equation, replacing vertex-based bounding box calculation. The goal of the generalized stage is to quickly eliminate absolutely non-intersecting object pairs. The system uses a scan-prune algorithm or a multi-level mesh method to output a set of candidate pairs that may intersect.

[0043] Step S3: Traverse the hybrid bounding volume hierarchy structure tree to perform generalized phase collision detection and filter out a list of potential collision pairs.

[0044] like Figure 5 As shown, step S3 in this embodiment includes continuous collision detection logic, which includes the following sub-steps: Step S301: Based on the motion velocity vector of the previous frame, generate a scan bounding volume stretched along the motion path for the parameterized entity object.

[0045] Step S302: Detect the overlap between the bounding volume of the scan and the bounding volumes of other objects.

[0046] Step S303: Add the corresponding object combination to the potential collision pair list only when the scan bounding volumes overlap.

[0047] Step S4: Prioritize the list of potential collision pairs based on the collision urgency index, and perform precise collision calculations in the narrow stage in sequence. For parameterized entity objects, the analytical geometry solution method is used, and for discrete mesh objects, the primitive intersection test method is used. Step S5: Output the collision detection results and update the physical state of the 3D scene data based on the collision detection results.

[0048] In step S4, the collision urgency index is used to quantify the calculation priority of each potential collision pair, and its calculation formula is as follows: ,in, This represents the collision urgency index; the higher the index, the higher the calculation priority. This represents the volume of the overlapping region of the axis-aligned bounding boxes of the two objects in a potential collision pair; and These represent the geometric center coordinate vectors of the two objects in a potential collision pair; Represents the square operation of Euclidean distance; To prevent the denominator from being zero, a preset minimum distance constant is used; and These represent the geometric computational complexity coefficients for the two objects. If the object is marked as a parameterized entity object, the coefficient takes a preset low value; if the object is marked as a discrete mesh object, the coefficient takes a preset high value. and The preset weighted balance factor; This represents the global normalization constant.

[0049] In this embodiment, the spatial item It describes the physical probability of collisions. Molecules. This represents the overlapping volume of the bounding boxes of two objects. The greater the overlap, the higher the probability of an actual collision. (Denominator) This represents the distance between the centers of two objects; the closer the distance, the greater the urgency. A minimum constant is introduced. This is to prevent division by zero errors when the centers of two objects coincide. This ensures that deeply intersecting object pairs are processed first.

[0050] Cost items A prediction of computational costs for different geometric types is introduced. For parametric solids, the computation is extremely fast due to the use of analytical solutions, therefore... It can be set to a small value, such as 1.0; for discrete meshes, computation is expensive due to the large number of patch traversal tests involved, therefore... It can be set to a large value, such as 10.0. Weighting factor The value is typically between 0.5 and 2.0, and can be dynamically adjusted according to the importance of the object.

[0051] Finally, the global normalization constant This is used to map exponents at different scene scales to a unified standard range, such as 0 to 100, so that the scheduler can make threshold determinations.

[0052] Through this sorting, the system can ensure that within the limited computation time window of each frame, the tasks most likely to collide and with the highest computational cost-effectiveness are prioritized. For tasks with low urgency or extremely high computational cost, they can be selectively discarded or postponed to the next frame when time is insufficient, thereby ensuring the stability of the frame rate.

[0053] like Figure 3As shown, in step S4 of this embodiment, when both objects in a potential collision pair are parameterized entity objects, the precise collision calculation in the narrow phase includes the following sub-steps: Step S401: Read the local coordinate system transformation matrices of the two parameterized entity objects.

[0054] Step S402: Transform the algebraic equation of one parameterized entity object into the local coordinate system of the other parameterized entity object; Step S403: Construct the analytical equation for the minimum distance between two parameterized entity objects; Step S404: Solve for the root of the minimum distance analytical equation using Newton's iteration method. If the root of the minimum distance analytical equation is less than or equal to zero, then a collision is determined to have occurred.

[0055] In step S4, when a potential collision pair includes a parameterized entity object and a discrete mesh object, the precise collision calculation in the narrow stage includes: obtaining a set of local facets in the discrete mesh object that are within the bounding box of the parameterized entity object; calculating the distance from each triangular facet in the set of local facets to the geometric center of the parameterized entity object; and determining whether the distance is less than the surface-defined threshold of the parameterized entity object to determine the collision penetration depth.

[0056] The specific execution logic of step S4 is based on the collision urgency index. Dynamic scheduling is implemented, including: establishing high-priority queues and regular-priority queues; and adjusting the collision urgency index. Potential collision pairs exceeding a preset urgency threshold are pushed into a high-priority queue, and the remaining potential collision pairs are pushed into a normal-priority queue. For tasks in the high-priority queue, main processor core resources are allocated to prioritize the precise collision calculation during the narrow phase. For tasks in the normal-priority queue, background threads or idle computation cycles are allocated to perform delayed calculations. When a potential collision pair contains at least one parameterized entity object and the collision urgency index... When the value is extremely high, the contact point of the previous time step is directly used as the initial value for hot start to perform fast convergence calculation.

[0057] In step S4, when using the analytical geometry solution method for parameterized entity objects, a caching mechanism is introduced, including: using the relative pose matrix of two parameterized entity objects as hash keys; querying the historical collision cache table, and if a corresponding hash key exists, directly reading the distance calculation result of the previous time step as the initial value; and performing a small number of iterative correction calculations based on the initial value to obtain the current collision state.

[0058] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for accelerating collision calculation using parametric geometry combined with a collision algorithm, characterized in that, The method includes the following steps: Step S1: Obtain the 3D scene data to be collision detected, and traverse all geometric objects in the 3D scene data. Classify and label the geometric objects as parametric solid objects or discrete mesh objects according to their geometric features. Step S2: Construct a hybrid bounding volume hierarchical structure tree, wherein the leaf nodes of the hybrid bounding volume hierarchical structure tree store the algebraic equation parameters of the parameterized entity object and the vertex data of the discrete mesh object, respectively. Step S3: Traverse the hybrid bounding volume hierarchical structure tree to perform generalized stage collision detection and filter out a list of potential collision pairs; Step S4: Prioritize the potential collision pair list based on the collision urgency index, and perform precise collision calculations in the narrow stage in sequence. The analytical geometry solution method is used for the parameterized entity object, and the primitive intersection test method is used for the discrete mesh object. Step S5: Output the collision detection results and update the physical state of the 3D scene data based on the collision detection results.

2. The collision calculation acceleration method combining parametric geometry and collision algorithm according to claim 1, characterized in that: In step S1, classifying and labeling the geometric object as a parametric solid object or a discrete mesh object based on its geometric features includes the following sub-steps: Step S101: Detect whether the shape topological features of the current geometric object match the preset basic body library, wherein the basic body library includes at least spheres, cuboids, cylinders and cones; Step S102: If the matching is successful, the geometric center, semi-axis length and rotation quaternion of the geometric object are extracted as parameters of the algebraic equation, and the geometric object is marked as the parameterized entity object; Step S103: If the matching fails, obtain the face index and vertex array of the geometric object, and mark the geometric object as the discrete mesh object.

3. A method for accelerating collision calculation by combining parametric geometry with a collision algorithm, characterized in that: In step S4, the collision urgency index is used to quantify the calculation priority of each potential collision pair, and its calculation formula is as follows: ,in, This represents the collision urgency index; the higher the index, the higher the calculation priority. This represents the volume of the overlapping region of the axis-aligned bounding boxes of the two objects in the potential collision pair; and These represent the geometric center coordinate vectors of the two objects in the potential collision pair, respectively. Represents the square operation of Euclidean distance; To prevent the denominator from being zero, a preset minimum distance constant is used; and These represent the geometric computational complexity coefficients corresponding to the two objects. If the object is marked as the parameterized entity object, the coefficient takes a preset low value; if the object is marked as the discrete mesh object, the coefficient takes a preset high value. and The preset weighted balance factor; This represents the global normalization constant.

4. A method for accelerating collision calculation by combining parametric geometry with a collision algorithm, characterized in that: In step S4, when both objects in the potential collision pair are the parameterized entity objects, the precise collision calculation in the narrow phase includes the following sub-steps: Step S401: Read the local coordinate system transformation matrices of the two parameterized entity objects; Step S402: Transform the algebraic equation of one of the parameterized entity objects into the local coordinate system of the other parameterized entity object; Step S403: Construct the minimum distance analytical equation between the two parameterized entity objects; Step S404: Solve for the root of the minimum distance analytical equation using Newton's iteration method. If the root of the minimum distance analytical equation is less than or equal to zero, then a collision is determined to have occurred.

5. A method for accelerating collision calculation using parametric geometry combined with a collision algorithm, characterized in that, Step S2 includes the following sub-steps: Step S201: Construct the hybrid bounding volume hierarchical structure tree in a top-down manner; Step S202: During the node splitting process, geometric objects that are spatially adjacent and belong to the same parameterized entity object are preferentially aggregated to the same parent node; Step S203: Calculate the bounding box for each node. If the node contains the parameterized entity object, generate a compact bounding box directly based on the algebraic equation parameters to replace vertex-based bounding box calculation.

6. A method for accelerating collision calculation using parametric geometry combined with a collision algorithm, characterized in that, In step S4, when the potential collision pair includes one parameterized entity object and one discrete mesh object, the precise collision calculation in the narrow stage includes: obtaining a set of local facets in the discrete mesh object that are located within the bounding box of the parameterized entity object; calculating the distance from each triangular facet in the set of local facets to the geometric center of the parameterized entity object; and determining whether the distance is less than a surface-defined threshold of the parameterized entity object to determine the collision penetration depth.

7. A method for accelerating collision calculation using parametric geometry combined with a collision algorithm, characterized in that, The specific execution logic of step S4 is based on the collision urgency index. Dynamic scheduling includes: establishing high-priority queues and ordinary-priority queues; and setting the collision urgency index... Potential collision pairs exceeding a preset urgency threshold are pushed into the high-priority queue, and the remaining potential collision pairs are pushed into the normal-priority queue. For tasks in the high-priority queue, main processor core resources are allocated to prioritize the execution of the narrow-stage collision precision calculation. For tasks in the normal-priority queue, background threads or idle computing cycles are allocated to perform delayed calculations. When a potential collision pair contains at least one of the parameterized entity objects and the collision urgency index... When the value is extremely high, the contact point of the previous time step is directly used as the initial value for hot start to perform fast convergence calculation.

8. A method for accelerating collision calculation using parametric geometry combined with a collision algorithm, characterized in that, Step S3 includes continuous collision detection logic, which includes the following sub-steps: Step S301: Based on the motion velocity vector of the previous frame, generate a scan bounding volume stretched along the motion path for the parameterized entity object; Step S302: Detect the overlap between the scanned bounding volume and the bounding volumes of other objects; Step S303: Add the corresponding object combination to the potential collision pair list only when the scan bounding volumes overlap.

9. A method for accelerating collision calculation using parametric geometry combined with a collision algorithm, characterized in that, In step S4, when using the analytical geometry solution method for the parameterized entity object, a caching mechanism is introduced, including: using the relative pose matrix of the two parameterized entity objects as hash keys; querying the historical collision cache table, and if a corresponding hash key exists, directly reading the distance calculation result of the previous time step as the initial value; and performing a small number of iterative correction calculations based on the initial value to obtain the current collision state.

10. Step S1 further includes preprocessing operations on the hybrid geometry, including: Identify target geometric objects that are composed of complex discrete meshes but whose overall outline approximates standard geometry; fit the target geometric object with the least squares method to obtain the maximum inscribed parameterized entity and the minimum circumscribed parameterized entity; in step S4, the maximum inscribed parameterized entity and the minimum circumscribed parameterized entity are used first for fast elimination test, and the original discrete mesh data is called only when the test results are unclear.

Images