Velocity estimation algorithms for audio-haptic simulations involving stick-slip.

Area of Science:

• Robotics
• Haptics
• Acoustic Simulation

Background:

• Real-time friction models require accurate velocity estimation from position data.
• Noisy position measurements, especially at high sampling rates, challenge velocity estimation.
• Poor velocity estimation degrades audio-haptic simulations, causing unnatural sounds and tactile sensations.

Purpose of the Study:

• To evaluate and compare various velocity estimation techniques for audio-haptic simulations.
• To identify the optimal method for balancing noise rejection and delay in velocity estimation.
• To assess the subjective quality of haptic feedback based on different estimation methods.

Main Methods:

• Thirteen distinct velocity estimation methods were automatically optimized and evaluated.
• A subset of optimized methods underwent subjective comparison.
• Performance was assessed based on accuracy, noise sensitivity, and delay in real-time friction modeling.

Main Results:

• No single velocity estimation method proved optimal across all gain levels.
• A sliding-mode differentiator feeding into a Kalman integrator demonstrated the best overall performance.
• Subjective evaluations confirmed the effectiveness of the chosen method in improving haptic and acoustic quality.

Conclusions:

• Velocity estimation remains a critical challenge in real-time haptic systems.
• The sliding-mode differentiator and Kalman integrator combination offers a robust solution for audio-haptic simulations.
• Physical velocity transduction sensors should be prioritized in haptic device design for superior performance.