Autonomous Vehicle Map Matching With Coalesced Cache Access
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The speed of location determination in vehicle navigation systems is limited by memory access speeds, rather than processing speed, due to the bottleneck in processing three-dimensional dynamic map data and static map data together.
Innovation Solution
A method involving obtaining and processing slices of static map data and frames of dynamic LIDAR data in a parallel processor's cache memory, arranging data in contiguous memory spaces for efficient access, and selecting the highest-scoring scan match candidate for vehicle positioning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If conventional memory access methods are used for processing three-dimensional dynamic map data and static map data, then data can be accessed, but the processing speed is limited by memory access speeds creating a bottleneck
Solution Approach 1:
The patent segments the three-dimensional map data into multiple two-dimensional slices at different heights. Each slice is processed independently and stored in separate memory locations, allowing parallel processing of multiple slices simultaneously. This segmentation transforms the bottleneck of accessing large 3D data into parallel access of smaller 2D slices, significantly improving processing speed and reducing data latency.
2Reliability
If all blocks of static map data are loaded into cache memory, then complete data is available for processing, but memory bandwidth is consumed and processing time increases
Solution Approach 1:
The patent implements selective loading of only necessary map data blocks into cache memory based on the vehicle's current position and predicted movement. Instead of loading all static map data, the system calculates which blocks are likely to be needed and loads only those. This partial loading approach maintains data reliability for current processing needs while significantly reducing memory bandwidth consumption and improving processing throughput.
Solution Approach 2:
The system performs preliminary calculations to predict which map blocks will be needed based on the vehicle's current position and trajectory. By pre-identifying and loading only the necessary blocks before they are actually needed for processing, the system ensures data availability without consuming excessive memory bandwidth, thus maintaining reliability while improving productivity.
Data Source
AI summary
Systems, methods, and non-transitory computer readable media are provided for obtaining a slice of static map data comprising a plurality of blocks, each block comprising a plurality of cells, each, each cell having a cell value indicating a probability that an object is present in the cell; loading the slice into a cache memory of a parallel processor; arranging the static map data in the cache memory in contiguous memory spaces assigned to a group of workers of the parallel processor that have coalescing constraints; loading a frame of dynamic map data into the cache memory; obtaining a plurality of scan match candidates each representing a possible position and attitude of the vehicle; processing, in the parallel processor, the static and dynamic map data and the candidates to generate results each representing a candidate and score; and selecting the candidate having the highest score as a vehicle position.