Adaptive Non-linear Filter for Ultrasonic Object Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Ultrasonic sensors face sensitivity losses when detecting objects with high amplitude reflections, such as bushes, which are difficult to distinguish from noise, especially at longer distances, leading to inaccurate object recognition.
Innovation Solution
An adaptive non-linear filter structure with a variable time constant is implemented, using the difference between filtered and detected signals, as well as the average amplitude profile, to enhance sensitivity and separate useful signals from noise, allowing for improved object recognition without the need for expensive sensor components.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If an adaptive threshold is used for object detection, then the detection sensitivity is improved, but sensitivity losses occur in the vicinity of strongly reflecting objects
Solution Approach 1:
The patent applies a dynamic filtering approach where the filter characteristics are continuously adapted based on the local signal environment. The filter transitions from a static adaptive threshold to a dynamic system that adjusts its behavior according to the detected signal properties, allowing it to maintain high sensitivity while adapting to varying reflection conditions.
Solution Approach 2:
The invention changes the parameter of the filter by introducing a variable time constant that adapts based on the local amplitude profile. This parameter change allows the filter to optimize its performance for different signal conditions, particularly improving detection near strong reflectors by adjusting the filtering characteristics to match the local environment.
2Object-affected harmful factors
If a standard linear low-pass filter is used, then noise is reduced, but signal delay increases and position accuracy decreases
Solution Approach 1:
The patent replaces the static linear low-pass filter with a dynamic non-linear filter whose time constant varies based on the local signal characteristics. This dynamic adaptation allows the filter to provide strong noise reduction when needed while minimizing signal delay by reducing the time constant in regions where position accuracy is critical.
Solution Approach 2:
The invention changes the time constant parameter of the low-pass filter from a fixed value to a variable parameter that adapts to local conditions. By modifying this parameter based on the amplitude profile and signal characteristics, the filter achieves optimal balance between noise reduction and position accuracy in different spatial regions.
3Measurement precision
If expensive sensor components are used to increase sensitivity, then detection capability is improved, but device cost increases
Solution Approach 1:
The patent achieves improved detection sensitivity through parameter changes in the signal processing algorithm rather than through expensive hardware modifications. By adapting the filter time constant and threshold characteristics based on the local signal environment, the system achieves high detection capability using standard sensor components.
Solution Approach 2:
The invention substitutes mechanical/hardware solutions (expensive sensor components) with a software/algorithmic solution (adaptive non-linear filtering). This replacement achieves the same sensitivity improvement through intelligent signal processing rather than through costly physical modifications to the sensor itself.
Data Source
AI summary
An object recognition device for vehicles, including a sensor, a filter device and an analysis device. Signals from an object are detectable by the sensor. The detected signals are filterable by the filter device and then feedable to the analysis device. The analysis device is configured to recognize the object by analyzing the filtered signals. The filter device includes an averaging unit, which averages the profile of the amplitudes of the detected signals, and a non-linear filter unit, which is provided based on a linear low-pass filter with a variable time constant and filters the detected signals. A value profile corresponding to the difference between the averaged amplitude profile and the filtered signals is feedable to the analysis device for analyzing and recognizing the object. The variable time constant is establishable based on the difference between the filtered signals and the detected signals and based on the averaged amplitude profile.