Multi-Sensor Occupancy Grid Fusion for Low-Latency Perception
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating occupancy grids in autonomous vehicles require substantial processing power, leading to increased latency and inefficiency, especially when dealing with multiple sensors and dynamic environments.
Innovation Solution
The proposed method involves obtaining detection information from a plurality of heterogeneous sensors, generating a single measurement grid by combining this information, determining occupancy probabilities for cells in the grid, and outputting an occupancy grid based on these probabilities, thereby reducing processing cycles and latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple sensors and complex processing algorithms are used to improve object detection accuracy, then detection precision is improved, but processing time and computational complexity increase
Solution Approach 1:
The patent segments the occupancy grid generation process into distinct phases: sensor data acquisition, measurement grid generation, occupancy probability computation, and grid output. By processing different sensor types (radar, lidar, camera) independently and generating measurement grids for each sensor separately before combining them, the system reduces computational complexity while maintaining detection accuracy across multiple sensor modalities
Solution Approach 2:
The patent performs preliminary processing by generating measurement grids for each sensor type before combining them into a final occupancy grid. This preliminary action allows the system to prepare sensor data in advance, compute occupancy probabilities for individual sensors, and then efficiently merge the results, thereby reducing overall processing latency while maintaining comprehensive object detection capability
2Measurement precision
If multiple sensors and complex processing algorithms are used to improve object detection accuracy, then detection precision is improved, but device complexity increases
Solution Approach 1:
The patent divides the complex multi-sensor processing system into modular components: individual measurement grid generators for each sensor type, separate occupancy probability computers, and a final grid combiner. This segmentation allows each component to be optimized independently and simplifies the overall system architecture while maintaining high detection accuracy through comprehensive multi-sensor fusion
Solution Approach 2:
The patent implements a universal occupancy grid generation framework that can process multiple sensor types (radar, lidar, camera) using the same fundamental algorithmic approach. The measurement grid generation and occupancy probability computation methods are designed to be sensor-agnostic, allowing the system to handle heterogeneous sensors through a unified processing pipeline, thereby reducing system complexity
Data Source
AI summary
Techniques are provide for generating occupancy grids based on inputs from multiple heterogeneous sensors. An example method for generating an occupancy grid includes obtaining detection information from a plurality of heterogeneous sensors, generating a single measurement grid based on the detection information from the plurality of heterogeneous sensors, determining occupancy probabilities for a plurality of cells in the single measurement grid, and outputting the occupancy grid based at least in part on the occupancy probabilities.