Analog Neural Network Circuit for Low-Power Tire Sensor Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current vehicle data management systems face inefficiencies in processing and transmitting sensor data, leading to high power consumption and bandwidth usage, which can be addressed by implementing analog hardware realization of neural networks for on-vehicle data processing.
Innovation Solution
A neural network circuit is placed near vehicle sensors to process sensor data and generate reduced-volume output data items, which are then wirelessly transmitted to the ECU, reducing the need for raw data transmission and conserving power and bandwidth.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If sensor data is transmitted wirelessly to the ECU for processing, then data communication is achieved, but power consumption increases substantially
Solution Approach 1:
The system segments the data processing function by placing a neural network circuit directly at the sensor node to perform local processing of sensor data, extracting only relevant features and transmitting compressed results to the ECU, thereby reducing wireless transmission power consumption while maintaining data communication reliability
Solution Approach 2:
The neural network circuit performs preliminary processing of sensor data locally before transmission, extracting embeddings and reducing data dimensionality in advance, so that only essential processed data needs to be transmitted wirelessly, significantly reducing the power consumption of wireless communication
2Loss of information
If raw sensor data with large volume is transmitted, then complete information is provided to the ECU, but bandwidth consumption increases
Solution Approach 1:
The neural network circuit extracts only the essential features and relevant information from raw sensor data by generating embeddings, removing redundant data while preserving critical information about vehicle conditions, thereby reducing transmitted data volume without causing information loss
Solution Approach 2:
The system changes the parameter representation of sensor data by transforming raw high-dimensional sensor readings into compressed low-dimensional embeddings through neural network processing, maintaining the essential information content while dramatically reducing the quantity of data that needs to be transmitted
3Productivity
If sensor nodes operate continuously at high sampling rates, then comprehensive monitoring is achieved, but power consumption increases
Solution Approach 1:
The system uses partial action by having the neural network circuit process only the essential features from sensor data rather than transmitting all raw data, allowing continuous monitoring at high sampling rates while reducing power consumption through selective processing and compressed transmission of only critical information
Data Source
AI summary
Systems, devices, integrated circuits, and methods are directed to on-vehicle data processing using analog hardware realization of neural networks. A vehicle obtains a temporal sequence of sensor data samples that is collected by a sensor system including a tire pressure sensor and/or a three-axis accelerometer. The sensor system is physically coupled to a tire of a vehicle. The temporal sequence of sensor data samples is converted into a plurality of first parallel data items, which is applied as a plurality of first inputs to a neural network circuit. The neural network circuit generates one or more output data items based on the plurality of first parallel data items. The one or more output data items indicate a condition of the road, the vehicle, or a component of the vehicle.