UltraSense uses machine learning to create better touch controls in cars

View all on-demand sessions from the Intelligent Security Summit here.


UltraSense Systems announced that it has created new touch control sensors for cars that are more accurate as they use machine learning.

The new In-Plane sensing automotive technology has the ability to enable multi-mode sensing and human-machine interface (HMI) control in the plane of the SmartSurface (or A-Surface). This drastically reduces the size and weight of the sensors, reducing the number of parts and complexity. It also allows modern designs and configurations.

And it helps deal with one of the worst things about touch controls: accidental activations. No one wants to tap a screen and miss the button they’re trying to press, or make a phone call from their pocket.

This translates into advantages in terms of sustainability and recyclability. It extends driving range and enables modern designs and new user experiences not previously possible, such as assistive controls for retractable steering wheels that require elegant, sleek form factors, the company said.

Event

Intelligent Security Summit on demand

Learn the critical role of AI and ML in cybersecurity and industry-specific case studies. Check out on-demand sessions today.

Look here

San Jose, Calif.-based UltraSense said In-Plane sensing is an important step toward being able to provide a full HMI experience by enabling the thinnest possible space. More than a capacitive ITO (Indium Tin Oxide) layer, this is defined by offering sensor fusion and enabling multimode sensing, processing and algorithms, feedback control: lighting, audio, haptics and secure connectivity.

UltraSense said this is a recipe for transformational change in reducing the existing car module depth. Combined with the TouchPoint family of HMI controllers, InPlane sensing enables designs that support all types of Smart Surface HMI interactions through the widest range of materials, in addition to capacitive plastic and glass, the company said. Smart Surfaces can now work through natural materials such as wood and leather to metal and other soft surfaces.

UltraSense Systems uses machine learning to perform more accurate touch detection.

“Traditional interface modules were measured in inches of thickness, with In-plane sensing we have it
approximately millimeters of surface thickness, with full solid-state HMI controller capabilities, from multimodal sensing and feedback control of lighting, audio and haptics,” Mo Maghsoudnia, CEO of UltraSense, said in a statement. “InPlane sensing principles combined with our TouchPoint family of HMI controllers deliver the thinnest HMI that works with the widest range of materials, and this technology is applicable for HMI experiences for automotive interior and exterior, industrial and consumer applications.”

TouchPoint HMI controllers provide multi-mode sensing (CapForce, UltraForce, TapForce), processing and AI machine learning algorithms, feedback control to drive lighting, audio, and haptics, and secure connectivity options.

UltraSense is a global company headquartered in Silicon Valley with offices in Taiwan, China, Korea, Japan and Europe. The company’s investors include Robert Bosch Ventures, Artiman Ventures, Abies Ventures, Sony Innovation, Sparx Group and Asahi Kasei.

UltraSense said its HMI controller delivers accuracy greater than one sense (i.e. capacitive only), eliminating “accidental activations.” The sensor can also sense through the widest range of materials, while the capacitive sensor can sense through plastic and glass. Car manufacturers want to distinguish themselves and see whether sensors can work with luxury materials such as metal and natural materials such as wood or leather.

I asked how the technique works. Much of the founding leadership team came from MEMS gyro/accelerometer leader InvenSense, which was acquired by TDK. Their MEMS technology powers most mobile phones today.

UltraSense responded that dual-mode detection using Cap and Force can significantly improve accuracy at a point of contact (usability score increases from 88% to 94%). With machine learning and local processing, the company can further improve accuracy (usability score rises above 96). % approaching 99%). This further removes “accidental activations”.

For example, steering wheel buttons for a driver holding the wheel at 10 o’clock and 2, or 9 o’clock and 3, will cover most conditions. But one automaker came across the “trucker’s attitude.” During long journeys, such drivers hold the steering wheel by the spokes and accidentally activate the buttons with their palms. Ultrasense ML technology can identify and remove that scenario of accidental activations.

The sensors also have integrated processing to deliver “zero latency” necessary for proper sensing-feedback interactions (think of a strange movie where the mouth and voice tracks don’t match). This is similar to the effect that occurs with in-car haptic response, where the MCU is a centralized MCU that multitasks and introduces random latency, resulting in double-clicking or pressing something not working. An integrated local processor eliminates this situation.

It also has feedback control of lighting, audio and haptic feedback, offering the best control without latency, the company said.

The company has less than 100 employees and was founded in 2018. UltraSense is integrated into consumer devices such as the LG Velvet and Rollable phones and with CASE (Connected, Autonomous, Shared Vehicle, Electrified) initiatives.

VentureBeat’s mission is to become a digital city plaza where tech decision makers can learn about transformative business technology and execute transactions. Discover our Briefings.