Sensor Simulation Object Detection via Tile-Based Spatial Indexing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current sensor simulation methods in vehicle technology face challenges in achieving real-time capability in complex traffic scenarios with numerous static and dynamic objects, as they require significant computing resources, leading to inefficiencies and non-real-time performance.
Innovation Solution
A method that updates sorted coordinate lists to identify objects within a pre-selection area, focusing only on objects entering or leaving the area, thereby reducing computational load and ensuring real-time processing by considering the appearing and disappearing sections of the pre-selection area.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all objects in the simulated world are checked to identify objects within the preselection area, then complete object detection is achieved, but computing capacity requirements increase proportionally with the number of objects
Solution Approach 1:
The simulated world is divided into multiple tiles, each managing its own subset of objects. This segmentation allows the system to process only relevant objects in each tile rather than checking all objects globally, reducing computational load while maintaining detection completeness within sensor range.
Solution Approach 2:
The patent implements a broad phase preselection using tile-based spatial indexing before the narrow phase detailed detection. Objects are preliminarily filtered by checking which tiles overlap with the sensor detection area, and only objects within those tiles are subjected to precise detection checks, significantly reducing the number of computations required.
2Productivity
If tile-based partitioning is used to reduce computational load, then processing efficiency improves, but real-time capability cannot be reliably ensured in dense traffic situations
Solution Approach 1:
The patent dynamically adjusts the detection strategy based on tile overlap with the sensor detection area. Tiles are actively managed and updated as the sensor moves, with the system adapting which objects require detailed checking. This dynamic approach ensures real-time capability is maintained even as traffic density and sensor position change.
Solution Approach 2:
The broad phase performs preliminary tile-based filtering to identify which tiles may contain detectable objects before proceeding to the narrow phase. This preliminary action reliably bounds the computational work required, ensuring real-time performance by preventing unnecessary processing of objects outside sensor range.
3Measurement precision
If the broad phase processes a large number of objects to ensure complete detection, then detection accuracy is maintained, but real-time performance deteriorates due to excessive computing capacity allocation
Solution Approach 1:
The patent extracts and processes only the relevant subset of objects that lie within or may intersect the sensor detection area. By using tile-based spatial indexing, the system extracts only those objects that could potentially be detected, discarding the vast majority of objects that are clearly outside the detection range, thus maintaining accuracy for relevant objects while dramatically reducing processing time.
Solution Approach 2:
The broad phase performs preliminary spatial filtering using tile overlap calculations to identify candidate objects before the narrow phase performs precise detection. This preliminary action maintains detection accuracy by ensuring all potentially detectable objects are considered, while reducing time loss by eliminating objects that are definitely outside sensor range.
Data Source
Figure 1
Figure 2
Figure 3a~3c
AI summary
The present invention relates to a computer-implemented method and a simulation system 50 for detecting objects 31, 32, 33, 110, 201 which are detected by a sensor 11, wherein the objects 31, 32, 33, 110, 201 and the sensor 11 move relative to each other in a simulated environment 100, 200. The procedure comprises step S100 of identifying objects 110, 201 that lie within a preselection area 30, 230, 231 in the simulated environment 100, 200, wherein the preselection area 30, 230, 231 includes a detection area 20 of the sensor 11, and step S200 of determining whether an object 31, 32, 33, 110, 201 in the preselection area 30, 230, 231 is detected by the sensor detection area 20.According to the invention, in step S100, objects 31, 32, 33, 110, 201 that are located in the preselection area 30, 230, 231 at a given time are identified by maintaining lists 210, 220 with sorted coordinates along dimensions of the simulated environment 100, 200. These lists identify such objects 31, 32, 33, 110, 201 that were located in the preselection area 30, 230, 231 at an earlier time according to step S110. The coordinate lists 210, 220 are updated in step S120 by adding or removing object identifications in step S124. The invention further relates to a corresponding sensor system 50 and a corresponding computer program 60.