Intermediate ML Input Split for Faster Object Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current techniques for processing sensor data in autonomous vehicles face challenges in resource-constrained environments due to the need to handle substantial input data, which can lead to processing restraints and inefficiencies.
Innovation Solution
The approach involves processing multi-channel image data using a two-part machine-learned model, where the first portion handles spatial features and the second portion handles non-spatial features, allowing for parallel processing and aggregation of results to determine classification probabilities, thereby optimizing resource usage and improving processing times.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If substantial input data is processed to determine object classifications, then classification accuracy is improved, but processing time and resource consumption increase
Solution Approach 1:
The patent divides the machine-learned model into two distinct portions: a first portion that processes spatial feature data and a second portion that processes non-spatial feature data. This segmentation allows parallel processing of different feature types, reducing overall processing time while maintaining comprehensive data analysis for accurate classification
2Measurement precision
If substantial input data is processed to determine object classifications, then classification accuracy is improved, but resource consumption increases
Solution Approach 1:
The model is segmented into two portions that can process different feature types in parallel, allowing for more efficient resource utilization. The first portion handles spatial features while the second portion handles non-spatial features, reducing redundant processing and lowering overall power consumption
Solution Approach 2:
The patent processes only the necessary portions of data through specialized model portions rather than processing all data uniformly. By directing spatial features to the first portion and non-spatial features to the second portion, the system avoids excessive processing of irrelevant data, optimizing resource usage
3Measurement precision
If comprehensive sensor data is processed, then detection accuracy at greater ranges is improved, but memory requirements increase
Solution Approach 1:
The patent segments the data processing into two distinct portions: spatial feature processing and non-spatial feature processing. This segmentation allows for optimized memory management where each portion can be designed to handle specific data types efficiently, reducing overall memory requirements while maintaining comprehensive analysis
Data Source
AI summary
Techniques for determining a classification probability of an object in an environment are discussed herein. Techniques may include analyzing sensor data associated with an environment from a perspective, such as a top-down perspective, using multi-channel data. From this perspective, techniques may determine channels of multi-channel input data and additional feature data. Channels corresponding to spatial features may be included in the multi-channel input data and data corresponding to non-spatial features may be included in the additional feature data. The multi-channel input data may be input to a first portion of a machine-learned (ML) model, and the additional feature data may be concatenated with intermediate output data from the first portion of the ML model, and input into a second portion of the ML model for subsequent processing and to determine the classification probabilities. Additionally, techniques may be performed on a multi-resolution voxel space representing the environment.