Intent Inference for Hand Pointing Gesture-Based Interactions in Vehicles.

Area of Science:

• Human-computer interaction
• Automotive technology
• Probabilistic modeling

Background:

• In-vehicle touchscreens demand significant visual and cognitive resources, distracting drivers.
• Driving conditions can cause perturbed user input, leading to errors and reduced system usability.

Purpose of the Study:

• To propose intent-aware displays that enhance in-vehicle interaction safety and efficiency.
• To reduce the time and cognitive effort required for in-vehicle selection tasks.

Main Methods:

• Utilizing a pointing gesture tracker and Bayesian destination inference algorithms.
• Modeling hand pointing gesture movements as a destination-reverting process.
• Implementing a Kalman filter-type prediction routine with low computational burden.

Main Results:

• High-confidence intent detection early in the pointing gesture.
• Significant reduction in time and effort for selection tasks.
• Demonstrated gains in an instrumented vehicle under various road conditions.

Conclusions:

• Intent-aware displays offer a safer and more efficient alternative to traditional touchscreens.
• The proposed probabilistic framework is computationally efficient and robust.
• This technology has the potential to significantly improve driver safety and user experience.