Ultrasonic Echo Feature Encoding for Low-Bandwidth Vehicle Buses
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for data transmission in ultrasonic systems for vehicle environment recognition exceeds the limited bandwidth of vehicle data buses, necessitating a method to enhance data compression without compromising recognition reliability and requiring additional bandwidth.
Innovation Solution
A method for transmitting ultrasonic data via a vehicle data bus that involves extracting predetermined signal profile characteristics from echo signals, allocating them to classes with identifiers, and transmitting only the identifiers and associated parameters, allowing reconstruction of the echo signal in a data processing device.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the quantity of data transmitted from the ultrasonic system to the data processing device is increased to improve obstacle recognition reliability, then the recognition accuracy is improved, but the data bus bandwidth is exceeded
Solution Approach 1:
The patent extracts only the essential signal profile characteristics (envelope signal, zero crossings, peaks, valleys) from the complete echo signal for transmission. By taking out only the relevant features needed for obstacle recognition rather than transmitting the entire signal, the data quantity is reduced while maintaining recognition reliability.
Solution Approach 2:
The patent creates a simplified representation (copy) of the echo signal by encoding signal profile characteristics into compact data structures with identifiers and parameters. This copied representation contains the essential information for obstacle recognition but occupies significantly less bandwidth than the original signal.
2Loss of information
If complete echo signals are transmitted to perform sensor fusion and object recognition in the computer system, then the information content is maximized, but the data bus load increases excessively
Solution Approach 1:
The patent extracts essential signal characteristics (envelope signal, zero crossings, peaks, valleys) from the complete echo signal. These extracted features preserve the critical information needed for obstacle recognition and sensor fusion while eliminating redundant data, thus reducing transmission quantity without significant information loss.
Solution Approach 2:
The patent transforms the continuous echo signal into discrete parameter representations (identifiers and numerical parameters describing signal profile characteristics). This parameterization converts high-volume continuous data into compact discrete data while retaining the essential information content for analysis.
3Quantity of substance
If data compression is applied to reduce transmission quantity, then the bandwidth usage is optimized, but the complexity of signal processing increases
Solution Approach 1:
The patent performs preliminary extraction of signal profile characteristics at the ultrasonic sensor level before transmission. By pre-processing the signal to identify and encode only the essential features (envelope, zero crossings, peaks, valleys), the complexity of processing is distributed to where the data is generated, simplifying the overall system architecture.
4Quantity of substance
If obstacle recognition is performed in the ultrasonic sensor system itself, then the data transmission quantity is reduced, but the recognition accuracy decreases due to limited processing capability
Solution Approach 1:
The patent extracts essential signal characteristics that preserve the information needed for accurate obstacle recognition. By transmitting these extracted features rather than complete signals or simple presence/absence data, the system enables accurate recognition at the data processing device while minimizing transmission requirements.
Data Source
AI summary
A method for transmitting data from an ultrasonic sensor to a computer system includes forming a feature vector signal from an electric reception signal; recognizing signal objects in the reception signal and classifying the signal objects according to predetermined signal object classes. The signal objects are forms or sequences of forms. At least one object parameter allocated to the signal object and one symbol for the signal object class are allocated to each signal object, or for each signal object, at least one signal object parameter and a symbol object are determined. The method further includes transmitting the symbol and the at least one signal object parameter to the computer system as data of a recognized signal object.