VentureIsrael News

NeuroSense Can Speed Motorcycle Airbag Deployment, Saving Lives

Portfolio Companies
The global motorcycle airbag market, valued at $410 million U.S. in 2023, is projected to reach $650 million by 2029, with a CAGR of 8.05%. This growth is fueled by rising awareness of rider safety and technological advancements. Motorcycles lack the protective structure of cars, increasing accident risks that lead to higher demand for safety features like airbags.

Motorcycle airbags, designed to mitigate injury by deploying during collisions, are becoming key safety components. These systems can be integrated into motorcycles or used as wearables, enhancing overall rider protection. Our focus here is on wearable airbags.

Wearable airbag systems use sensors to detect certain movements, such as sudden deceleration, changes in angle, or impacts, which indicate a crash or fall. There are different sensor types: mechanical tethers (a cord attached to the motorcycle that pulls when the rider separates from the bike) and electronic sensors (accelerometers and gyroscopes, sometimes combined with motorcycle sensors).

When the rider with the mechanical system is thrown off the bike, the rope triggers a gas canister that inflates the airbag. In electronic systems, advanced airbag solutions continuously analyze sensor data to detect crashes in milliseconds. An electronic trigger signals the inflation system when a dangerous event is detected, ensuring rapid deployment of the airbags.

Mechanical tether systems for wearable motorcycle airbags offer several advantages, primarily due to their affordability, simplicity, and lack of battery maintenance hassle. However, mechanical tether systems have some notable drawbacks. They activate only when the rider is physically separated from the motorcycle, which means they may not work at low-speed or low-impact crashes where the rider remains seated. Additionally, the response time can be slower than that of electronic systems. A rope can restrict the rider’s movement and pose a risk of accidental activation if the rope is not disconnected before dismounting.

Electronic systems for wearable motorcycle airbags offer significant advantages, mainly due to their ability to detect a broad range of crash scenarios and immediate response time. The airbag must inflate in less than 200 milliseconds from detecting an event. This timeframe includes detecting the incident, triggering the inflation system, and inflating the airbag with gas canisters, typically CO₂ or argon.

Their ability to activate even when the rider remains seated during an incident enhances safety. Electronic systems can be embedded in vests, jackets, and helmets to offer broader applicability.

However, electronic systems are more expensive than their mechanical counterparts. They require regular charging to function correctly, which can be inconvenient and leave the airbag inactive if the battery is not properly maintained. Furthermore, the complexity of electronic systems means there is a potential for technical issues, such as sensor errors or software malfunctions. A false positive, where the airbag inflates unnecessarily, can lead to inconvenience or potential injury while incurring additional costs for resetting or replacing components. Conversely, a false negative, where the airbag fails to deploy during a real incident, can lead to severe injuries or even fatalities.

To address the drawbacks of electronic systems, POLYN has an AI solution for personal wearable airbags.

From a cost perspective, POLYN offers the NASP (Neuromorphic Analog Signal Processing) NeuroSense chip, which features efficient neural network-based fall detection for low-cost, compact sensor nodes easily integrated into vests, jackets, and helmets.

Regarding power consumption, NeuroSense significantly reduces the number of sensor node charging cycles because the NASP chip operates at the microwatt level. While current solutions typically offer a battery duration of 24 to 30 hours, NeuroSense allows sensor nodes to work for up to several months, depending on the battery type.

Due to its remarkably low latency, NeuroSense has significant potential to improve wearable airbag systems’ response time. Current electronic airbag solutions, which rely on microcontroller units (MCUs), have a triggering latency (from fall detection to the decision to inflate) of around 80 milliseconds in the best cases. In contrast, NeuroSense offers a breakthrough NASP processing taking around 50 microseconds. The chip doesn’t need data conversion to digital when applied with analog sensors. Reducing latency greatly enhances the effectiveness of wearable airbags and provides more efficient injury protection.

To minimize activation errors, POLYN leverages the latest advancements in machine learning to train neural network models capable of significantly reducing false positives and negatives. A key factor in achieving this is using labeled data to train the models. POLYN is actively seeking collaborations to develop neural networks compatible with its NASP architecture and is also prepared to adapt existing models for integration into its NASP chips.

POLYN’s expertise in sensor integration and ability to adapt and optimize models for specific application use cases positions it as a key player in developing highly efficient, reliable electronic airbag systems for rider safety.