Robotic Arm Acoustic Collision Localization Using Calibrated Onset Times
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Robotic devices face challenges in accurately localizing collisions due to dispersive plate wave propagation and structural heterogeneity, which complicates time-of-arrival-based methods and triangle-based localization algorithms, especially when microphones are spaced along a robotic arm.
Innovation Solution
A method that calibrates microphones' P-wave time-of-arrival (TDoA) into a one-dimensional manifold using a scoring function to estimate collision location by generating a virtual onset time set based on primary onset times from evenly spaced marker locations, filtering audio signals to remove noise, and determining scores based on standard deviation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If correlation based time distance of arrival (TDoA) methods are used for collision localization, then the method is simple to implement, but the method becomes infeasible due to dispersive plate wave propagation causing different frequency components to have different velocities
Solution Approach 1:
The patent transforms the localization problem from time-domain correlation to frequency-domain analysis. By identifying that plate waves are dispersive (velocity depends on frequency), the method changes the parameter used for localization from simple TDoA to frequency-specific velocity compensation. The scoring function evaluates multiple frequency components and synthesizes their results, resolving the contradiction by adapting the approach to the physical characteristics of plate wave propagation.
2Area of stationary object
If microphones are spaced along a robot arm to enable collision detection, then the coverage area is improved, but the structural heterogeneity and dispersive nature of plate waves make signal interpretation difficult
Solution Approach 1:
The patent introduces a scoring function that provides feedback on the quality of localization results. The function evaluates the consistency of signals across multiple microphones and frequency components, assigning scores that indicate confidence in the localization. This feedback mechanism helps resolve the difficulty of signal interpretation by quantifying the reliability of detected collisions and guiding the selection of valid detection results.
Solution Approach 2:
The patent performs preliminary calibration to establish the dispersive characteristics of plate waves in the specific robot structure. By pre-characterizing the velocity-frequency relationship for the robot's plate structures, the system prepares compensation data before actual collision detection. This preliminary action simplifies real-time signal interpretation by providing reference information about expected wave propagation behavior.
3Measurement precision
If plate wave propagation is used for collision detection, then the detection sensitivity is improved, but the dispersive nature causes different frequency components to arrive at different times making localization infeasible
Solution Approach 1:
The patent moves the localization problem from one-dimensional time-of-arrival measurement to a multi-dimensional solution space. Instead of relying on single-time TDoA, the method analyzes signals across multiple frequency dimensions and synthesizes localization results from the combined information. This dimensional expansion allows the system to compensate for time dispersion by utilizing the additional frequency domain information.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables accurate wireless collision localization on robotic devices by filtering noise and using a calibrated manifold to estimate collision location effectively, overcoming dispersive wave propagation and structural heterogeneity issues.
Implementation Method 1
Sound waves resulting from collisions that happen between robots and surrounding obstacles are impulsive in nature. These impulse waves are typically lamb waves (plate waves) because robots are generally constructed with solid plates.
Implementation Method 2
obtain audio signals from a plurality of acoustic sensors spaced apart along the robotic device
Implementation Method 3
filtering raw audio signals with a band stop filter if the robotic device is moving or passing the raw audio signals if the robotic device is static. A center frequency of the band stop filter may correspond to a fundamental frequency of a motor powering the robotic device.
Implementation Method 4
filtering the obtained audio signals to provide low pass audio signals and high pass audio signals for the obtained audio signals
Data Source
AI summary
A method of collision localization on a robotic device includes obtaining audio signals from a plurality of acoustic sensors spaced apart along the robotic device; identifying, based on a collision being detected, a strongest audio signal; identifying a primary onset time for an acoustic sensor producing the strongest audio signal, the primary onset time being a time at which waves propagating from the collision reach the acoustic sensor producing the strongest audio signal; generating a virtual onset time set, by shifting a calibration manifold, based on the identified primary onset time, the calibration manifold representing relative onset times from evenly spaced marker locations on the robotic device to the plurality of acoustic sensors; determining scores for the marker locations based a standard deviation of elements in the virtual onset time set; and estimating a location of the collision based on a highest score of the determined scores.