3D Inverse Sensor Model for Occupancy Grid False Detection Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Radar systems used in vehicle navigation face challenges with false detections and low angular resolution, leading to persistent uncertainties in occupancy grids, which result in incorrect tracking of moving objects.
Innovation Solution
A method employing a dual inverse sensor model, including a positive and negative inverse sensor model, to update likelihood values in an occupancy grid, where empty space likelihoods are assigned based on lack of detection, using logarithmic updates to reflect the current state of the environment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional inverse sensor model is used to create occupancy grid, then object detection capability is provided, but false detections persist and tracking accuracy deteriorates due to low angular resolution and large uncertainties
Solution Approach 1:
The patent segments the occupancy grid into multiple bins along angular directions (azimuth and elevation), allowing independent likelihood updates for each bin. This segmentation enables precise tracking of object positions by distributing probability mass across discrete angular bins, thereby reducing false detections while maintaining detection reliability.
Solution Approach 2:
The patent transforms the inverse sensor model by changing the parameter representation from continuous probability distributions to discrete likelihood values assigned to specific angular bins. This parameter transformation allows the system to handle angular uncertainties by distributing likelihoods across multiple bins rather than concentrating them, improving both detection reliability and angular precision.
2Productivity
If occupancy likelihoods are updated based on detections, then object tracking is enabled, but false detections perpetuate in the occupancy grid due to persistence of likelihoods in azimuth and elevation directions
Solution Approach 1:
The patent implements dynamic likelihood updates in the occupancy grid where likelihood values are continuously adjusted based on new detections and non-detections. The system dynamically redistributes probability mass across angular bins, allowing the occupancy grid to adapt to moving objects while reducing persistent false detections through time-varying likelihood assignments.
Solution Approach 2:
The patent incorporates feedback mechanisms where the occupancy grid is continuously updated based on comparisons between predicted object positions and actual sensor detections. When detections do not confirm expected object positions, the system provides negative feedback by reducing likelihoods in corresponding bins, thereby eliminating false detections while maintaining accurate tracking of genuine objects.
3Device complexity
If radar system operates with low angular resolution, then device complexity is reduced, but uncertainty in azimuth and elevation increases leading to poor tracking performance
Solution Approach 1:
The patent compensates for low angular resolution by introducing an additional dimension of information representation through multiple occupancy bins along both azimuth and elevation directions. This dimensional expansion allows the system to recover lost angular position information by distributing likelihoods across the bin grid, effectively increasing measurement precision without increasing radar hardware complexity.
Data Source
AI summary
A vehicle, navigation system of the vehicle, and method for observing an environment of the vehicle. The navigation system includes a sensor and a processor. The sensor obtains a detection from an object located in an environment of the vehicle at a with-detection direction with respect to the sensor that includes the detection. The processor determines a no-detection direction that does not include the detection, assigns an empty space likelihood to an occupancy grid of the sensor along the no-detection direction, and generates a map of the environment using the occupancy grid.