Insect-Inspired, Spike-Based, in-Sensor, and Night-Time Collision Detector Based on Atomically Thin and...

Area of Science:

• Optoelectronics
• Biomimetic sensors
• Autonomous systems

Background:

• Nighttime collision detection is challenging for autonomous vehicles due to limited visual data.
• Current technologies like LiDAR and image sensors are power-intensive and computationally expensive.
• Insects achieve efficient collision detection using minimal neural resources.

Purpose of the Study:

• To develop a simplified, energy-efficient collision detection system inspired by insect behavior.
• To leverage novel optoelectronic integrated circuits and memtransistor technology for enhanced sensing.
• To demonstrate a viable alternative to complex image-based systems for nighttime autonomous navigation.

Main Methods:

• Implementation of insect-inspired collision detection algorithms.
• Integration with in-sensor processing using atomically thin, photosensitive memtransistor technology.
• Testing under various real-life nighttime collision scenarios.

Main Results:

• The proposed system effectively detects vehicles on collision courses at night.
• Eliminates the need for traditional image capture and processing.
• Achieves a small footprint (∼40 μm²) and minimal energy consumption (few hundred picojoules).

Conclusions:

• Insect-inspired collision detection with memtransistor technology offers a simplified and efficient solution.
• This technology can significantly enhance the safety of autonomous vehicles, especially in low-light conditions.
• The system's low power and size make it suitable for augmenting existing autonomous vehicular sensors.