Sensor Data Processing via Point Cloud Feature Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sensor data processing methods for surroundings sensors, such as radar sensors, are inefficient as they either lose significant information when compressing data into point clouds or require large data transmission rates.
Innovation Solution
A method that processes sensor data from surroundings sensors by calculating point clouds and inputting these points along with their associated information sections into a trained processing model to generate feature vectors, which are then output for further processing, reducing data size and transmission requirements while maintaining information content.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If sensor data is compressed into point clouds, then data size is reduced, but information content is lost
Solution Approach 1:
The patent introduces an intermediate representation that includes both point cloud data and associated raw sensor data sections. This intermediary structure allows the system to maintain access to original information while working with compressed representations, effectively mediating between data compression and information preservation requirements.
Solution Approach 2:
The patent implements a nested data structure where point cloud points are embedded within a larger structure that also contains associated raw sensor data sections. This nesting allows multiple levels of data representation to coexist, enabling the system to use compressed data when sufficient while preserving access to full-resolution data when needed.
2Loss of information
If all sensor data is transmitted for further processing, then information content is maintained, but data transmission rate increases
Solution Approach 1:
The patent extracts and transmits only the essential components for further processing: point cloud points paired with their associated raw sensor data sections. By selecting and transmitting only these specific elements rather than all sensor data, the system reduces transmission requirements while maintaining the information necessary for accurate processing.
3Productivity
If only point cloud data is used for further processing, then data transmission is reduced, but processing accuracy decreases
Solution Approach 1:
The patent merges point cloud data with associated raw sensor data sections into a unified input structure for further processing. This combination allows processing models to leverage both the compressed spatial representation and the detailed original measurements, achieving accurate processing with reduced transmission requirements compared to sending all sensor data.
Data Source
AI summary
A method for sensor data processing of sensor data of a surroundings sensor. The method includes providing the sensor data of the surroundings sensor acquiring at least one target object in the surroundings of the sensor, calculating a point cloud including a spatial distribution of a plurality of points assigned to respective reflections of the target object from the sensor data, inputting the points and the information sections of the sensor data respectively assigned to them into a trained processing model, calculating feature vectors assigned the points by the processing model depending on the input, outputting the points and the assigned feature vectors for further processing by a further processing model. A processing device is also described.