Fused Occupancy Grid Steering for Reliable AV Trajectories
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face challenges in handling high-frequency, high-throughput compute requirements for processing sensor data from multiple offboard perception sensors in automated vehicles, particularly in ensuring secure and efficient trajectory calculations.
Innovation Solution
A scalable architecture is employed, utilizing multiple processing systems to pre-fuse occupancy grids from subsets of perception sensors before merging them at a further processing system, with timestamp management to ensure data freshness and redundancy for safe trajectory determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple offboard perception sensors are used to process sensor data at high frequency (5-20 hertz), then the reliability and coverage of trajectory determination is improved, but the computational load and bandwidth requirements increase significantly
Solution Approach 1:
The system divides the computational workload by assigning different subsets of sensor datasets to different processing systems. Each processing system handles a specific portion of the sensor data, processing it independently to generate occupancy grids. This segmentation reduces the computational burden on each individual system while maintaining overall system reliability through redundant processing paths.
Solution Approach 2:
The patent implements preliminary fusion of occupancy grids at each processing system before the final merging stage. By pre-processing and fusing occupancy grids locally at each processing system, the system reduces the amount of data that needs to be transmitted and processed at the central level, thereby reducing bandwidth requirements and computational complexity while ensuring data freshness through timestamp management.
2Speed
If sensor data is processed at high frequency (5-20 hertz) to ensure real-time trajectory updates, then the responsiveness of the automated vehicle is improved, but the energy consumption and computational resources increase
Solution Approach 1:
The computational task is divided into multiple parallel processing systems that each handle specific sensor datasets independently. This segmentation allows for efficient utilization of computational resources, enabling high-frequency processing (5-20 hertz) without concentrating all computational energy demands on a single system, thereby distributing energy consumption across multiple units.
Solution Approach 2:
By performing preliminary fusion of occupancy grids at each processing system before central merging, the system reduces the volume of data requiring high-speed transmission and processing. This pre-processing step maintains real-time responsiveness while reducing the overall computational energy required at the central processing level.
3Quantity of substance
If multiple processing systems are used to handle sensor datasets from distinct subsets of perception sensors, then the bandwidth requirements are reduced through distributed processing, but the system architecture becomes more complex
Solution Approach 1:
The system architecture is segmented into multiple processing systems, each handling distinct subsets of sensor datasets. This segmentation naturally reduces the bandwidth requirements for data transmission because each processing system works with a portion of the total data rather than all data centrally. The modular architecture manages complexity through standardized interfaces and protocols for fusion and merging operations.
Solution Approach 2:
The patent introduces an intermediary fusion step where occupancy grids from multiple sensors are combined at each processing system before being merged centrally. This intermediary fusion act as a mediator that reduces the data volume requiring transmission to the central system, thereby reducing bandwidth requirements while the standardized fusion protocol manages the architectural complexity.
4Reliability
If redundant processing systems are implemented for safety reasons, then the reliability and safety of trajectory computation is improved, but the computational overhead and resource requirements increase
Solution Approach 1:
Redundant processing systems are implemented by segmenting the computational workload across multiple independent processing units. Each system processes specific sensor datasets and generates occupancy grids independently. This segmentation provides inherent redundancy for safety-critical trajectory computation while maintaining computational efficiency through parallel processing, avoiding the overhead of redundant computation of the same data across all systems.
Solution Approach 2:
The preliminary fusion of occupancy grids at each processing system creates redundant but complementary computational paths. By fusing occupancy grids locally before central merging, the system ensures that redundant processing systems produce consistent and validated results, improving safety while optimizing computational efficiency by avoiding duplicate full-processing cycles.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention is notably directed to a method of steering an automated vehicle (2) in a designated area, thanks to a set (10) of offboard perception sensors (110 - 140). The method comprises repeatedly executing algorithmic iterations, where each iteration comprises the following steps. First, sensor data are dispatched to K processing systems (11, 12), whereby each processing system k of the K processing systems receives Nk datasets of the sensor data as obtained from Nk respective sensors of the set (10) of offboard perception sensors (110 - 140), where k = 1 to K, K ≥ 2, and Nk ≥ 2. The Nk datasets are subsequently processed at each processing system k to obtain Mk occupancy grids corresponding to perceptions from Mk respective sensors of the offboard perception sensors, respectively, where Nk ≥ Mk ≥ 1. The Mk occupancy grids overlap at least partly. Data from the Mk occupancy grids obtained are then fused, at each processing system k, to form a fused occupancy grid, whereby K fused occupancy grids are formed by the K processing systems (11, 12), respectively. The K fused occupancy grids are then forwarded to a further processing system (14), which merges the K fused occupancy grids to obtain a global occupancy grid for the designated area. Eventually, a trajectory is determined for the automated vehicle (2), based on the global occupancy grid. This trajectory is then forwarded to a drive-by-wire system (20) of the automated vehicle (2), to accordingly steer the latter. The invention is further directed to related systems and computer program products.