Radar SLAM Circuitry Using Particle Maps for Sensor-Free Pose Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current simultaneous localization and mapping (SLAM) techniques for autonomous mobile platforms face challenges in reliable pose estimation and mapping, especially in environments with poor GPS reception, and require additional sensors like IMUs, which increase costs and reduce system robustness.
Innovation Solution
A circuitry and method that utilize radar detection data to estimate ego-motion and update a set of particles, each containing the location, orientation, and occupancy grid map of a mobile platform, where the occupancy grid map represents the environment with cells assigned occupation probabilities, allowing for self-localization and mapping without additional sensors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If GPS and IMU sensors are used for pose estimation, then location accuracy is improved, but system cost and complexity increase
Solution Approach 1:
The patent extracts the essential function of pose estimation from the complex GPS+IMU sensor combination and implements it using only radar sensors. By taking out the unnecessary GPS and IMU components and relying solely on radar detection data processed through particle filtering, the system achieves pose estimation without increasing device complexity or cost.
Solution Approach 2:
The radar sensor performs multiple functions simultaneously - it provides both environmental detection data and ego-motion estimation. The system uses the radar's own detection data to estimate platform motion through particle filtering, making the radar sensor self-sufficient for pose estimation without requiring additional dedicated sensors.
2Reliability
If GPS and IMU sensors are combined for location estimation, then measurement reliability is improved, but device complexity increases
Solution Approach 1:
The patent removes the IMU sensor from the system architecture while maintaining measurement reliability. By using particle filtering to process radar detection data, the system extracts sufficient motion information without requiring the additional IMU hardware, thereby reducing system complexity while preserving reliability.
Solution Approach 2:
The patent replaces the mechanical inertial measurement system (IMU) with a signal processing approach (particle filtering) applied to radar data. This substitution eliminates the need for mechanical sensors while achieving comparable or superior reliability through computational methods.
3Measurement precision
If landmark extraction and matching is performed for SLAM, then mapping accuracy is improved, but computational cost increases
Solution Approach 1:
The patent extracts only the essential particles representing relevant environmental features from the full radar detection data, rather than performing comprehensive landmark extraction and matching. This selective extraction maintains mapping accuracy by focusing on key particles while significantly reducing computational processing requirements.
Solution Approach 2:
Instead of processing all detected landmarks and features, the patent applies partial action by selectively updating only the necessary particles in the particle filter that contribute to pose estimation and mapping. This approach achieves sufficient mapping accuracy without the excessive computational cost of processing all available data.
Data Source
AI summary
A circuitry for simultaneous localization and mapping for a mobile platform, wherein the circuitry is configured to:estimate, based on obtained radar detection data, an ego-motion of the mobile platform; andupdate, based on the estimated ego-motion and the obtained radar detection data, a set of particles, wherein each particle of the set of particles includes a location and an orientation of the mobile platform and an occupancy grid map that represents an environment of the mobile platform, wherein the occupancy grid map includes a plurality of cells and each cell of the plurality of cells is assigned an occupation probability which indicates a probability that the cell is occupied by a target.