Skip to main content
  1. Home
  2. Cars
  3. Emerging Tech
  4. News

This tiny sensor could help self-driving cars and robots see better in the dark

Penn State researchers have developed a light-adaptive photomemristor modeled on the human eye that achieves over 95% visual accuracy in shifting light conditions.

Add as a preferred source on Google
Waymo Jaguar I-PACE sensors close up
Waymo

Penn State researchers have developed a light-adaptive sensor component that could make autonomous vehicle cameras and robots far more reliable in shifting lighting conditions. The work, published Monday in Nature Communications, takes direct cues from how the human eye adjusts between bright and dark environments.

Biology as a blueprint

Current camera systems in self-driving cars are tuned for consistent lighting, which means accuracy drops when conditions shift rapidly, like moving from a dark road into oncoming headlights. The Penn State team, co-led by engineering professor Larry Cheng, looked to the eye’s rod and cone cell system for a solution. In the eye, rod cells contain pigments that bleach in bright light and gradually regenerate in darkness, allowing the eye to recalibrate its sensitivity continuously.

The researchers replicated that dynamic in a new type of photomemristor, a tiny sensor that captures light and converts it into electrical data. Their design uses two materials: a conductive gel-like polymer and titanium oxide. When light hits the titanium oxide, it generates a current that causes the polymer to absorb or release water depending on brightness, effectively self-regulating sensitivity in real time.

95% accuracy in mixed light

To test the design, the team built a 4×4 array of the sensors and paired it with a neural network, creating a basic machine vision system. They ran it through a variation of the standard eye chart test, asking it to identify an LED letter “F” against a backdrop tuned to fluctuating brightness levels. After seven training cycles, the system hit over 95% accuracy under mixed lighting conditions.

Recommended Videos

Each sensor measures just half a millimeter across, and Cheng says individual units can be connected in larger arrays to detect broader visual patterns without changing the physical size of each component.

Beyond autonomous vehicles, the team sees potential applications in factory robotics and, further down the line, assistive technology for people with visual impairments. A provisional patent has been filed, and next steps involve expanding the sensors into a multi-modal system capable of processing both visual and tactile data simultaneously.

The photomemristor joins a growing list of sensor innovations aimed at improving autonomous vehicle reliability, including a compact radar unit developed at Rice University earlier this year.

Pranob Mehrotra
Pranob is a seasoned tech journalist with over eight years of experience covering consumer technology. His work has been…
Tesla’s hidden door problem might end up fixing every car in America
This could permanently change how emergency exits are designed in American passenger vehicles.
Tesla Model Y Long Wheelbase Featured

Tesla's electronic door handles have been a safety talking point for years, mostly on Reddit threads and EV forums where people debate whether sleek design is worth the tradeoff when a car is on fire and power is out. 

Now the US government has formally weighed in (via The Washington Post). The outcome, I’d say, is bigger than most people think.

Read more
Ford’s UEV platform will feature Apple Maps built in
The integration will run through Apple's new MapKit for Automotive SDK and skip CarPlay's phone-mirroring approach.
Car, Transportation, Vehicle

Ford's upcoming budget electric vehicle won't force you to choose between a built-in map and one borrowed from your phone. Apple announced today that Ford's Universal Electric Vehicle (UEV) platform will ship with Apple Maps running natively on the car's screens. The feature will arrive in 2027 through Apple's new MapKit for Automotive SDK, a toolkit that lets automakers build Apple Maps straight into their dashboards.

An infotainment system that doesn't need your phone plugged in

Read more
WhatsApp’s new in-car experience is built around hands-free control
CarPlay and Android Auto users can now hear and reply to messages, view call history, and reach favorite contacts without touching their phone.
WhatsApp new CarPlay experience featured

Meta is overhauling how WhatsApp works in the car, giving CarPlay and Android Auto users a lot more to do than just sending texts and making calls. The company announced the change in a newsroom post today, alongside several other updates rolling out to the messaging app.

WhatsApp finally works hands-free in the car

Read more