Accuracy of estimating time to collision using binocular and monocular information

Area of Science:

• Visual Perception
• Human Factors
• Robotics and Autonomous Systems

Background:

• Accurate estimation of time to collision (TTC) is vital for safe navigation and interaction with dynamic environments.
• Previous research suggests both binocular and monocular cues contribute to TTC perception, but their relative importance, especially for small targets, remains debated.

Purpose of the Study:

• To quantify the just-noticeable difference (JND) and absolute accuracy in TTC estimation using binocular, monocular, and combined visual information.
• To determine the specific visual cues critical for accurate TTC estimation of small objects at close distances.

Main Methods:

• Participants estimated time to collision (TTC) with approaching targets of varying sizes (0.03 deg and 0.7 deg).
• TTC estimation accuracy and JND were measured under three conditions: binocular vision only, monocular vision only, and combined binocular and monocular vision.
• Data analysis focused on comparing JND and error percentages across different visual conditions and target sizes.

Main Results:

• Binocular vision alone allowed reliable discrimination of TTC variations for both small and large targets.
• Monocular vision was insufficient for reliable TTC discrimination with small targets, though effective for larger ones.
• Combined binocular and monocular vision resulted in the lowest TTC estimation errors (1.3–2.7%), approaching the accuracy needed for high-speed sports.

Conclusions:

• Accurate TTC estimation for small, nearby objects primarily relies on binocular information.
• Combined visual cues (binocular and monocular) offer the highest accuracy in TTC estimation, mimicking real-world conditions.
• The findings have implications for understanding human visual performance in dynamic environments and informing the design of autonomous systems.