A system for learning-based processing of radar data
By applying machine learning algorithms to process raw radar data and combining sensor data, the limitations of current radar systems are addressed, enhancing performance and robustness in vehicles.
JP2026522205APending Publication Date: 2026-07-07ZENDAR INC
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- ZENDAR INC
- Filing Date
- 2024-05-24
- Publication Date
- 2026-07-07
AI Technical Summary
Technical Problem
Current radar systems in vehicles process data independently, leading to limited performance and computational ambiguity, especially in harsh climatic conditions, and rely on assumptions that ignore the complexity of the real world, resulting in significant information loss.
Method used
Implementing machine learning algorithms, such as deep neural networks, to process raw radar data without calculating point clouds, combining data from multiple sensors, and using automated labeling techniques to enhance performance and robustness.
Benefits of technology
Improves radar system performance by accurately determining target characteristics and generating scene representations, overcoming limitations of conventional methods in challenging environments.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure 2026522205000001_ABST
Abstract
This disclosure provides a system for processing radar data. The system may comprise a vehicle located in an environment, a radar module associated with the vehicle, and an electronic processor configured to: receive an incoming radar signal from the radar module, including an indication of an object in the environment; process the incoming radar signal through one or more signal processing algorithms to determine a raw radar spectrum; process the raw radar spectrum through a machine learning computation model to determine a set of predicted environmental outputs; determine a representative model of the environment based at least in part on the set of predicted environmental outputs; and provide the representative model of the environment to an autonomous driving system.
Need to check novelty before this filing date? Find Prior Art