StradVision

Agile ADAS on Lean Hardware

Follow StradVision on :

Junhwan Kim, CEO, StradVisionJunhwan Kim, CEO
The automotive industry has upped the ante of digital innovation to develop safety-first autonomous driving and assistance capabilities. With significant investments in cutting-edge sensor technologies, many automotive companies are on the lookout for smart and capable ADAS (Advanced Driver-Assistance Systems) tools to effectively leverage sensors and enable vehicle intelligence. Junhwan Kim, CEO of StradVision, says: “Automotive OEMs and suppliers are pushing for connected ADAS that can provide real-time feedback through graphical image processing.”

StradVision is a solutions company that develops camera-based software technology to equip automotive manufacturers with robust and flexible ADAS.

“Our solution leverages deep neural networks (DNN) and artificial intelligence (AI) systems to enable vehicles with real-time detection of objects, pedestrians and vehicles for enhanced safety,” says Kim.
Utilizing the pre-existing hardware and camera sensors in the car, the solution detects obstacles in the driver’s blind spots and alerts the driver to potential accidents. Additionally, the level 2 ADAS solution prevents collisions by detecting lanes, abrupt lane changes, and vehicle speeds, even in poor lighting and weather conditions.

The solution’s ingenuity resides in its hardware-agnostic design. “We can seamlessly integrate our flexible software into a multitude of market available ADAS chipsets, enabling OEMs to save cost by avoiding the purchase of new sensors and hardware for compatibility,” says Kim. StradVision also provides the necessary chipsets and components to help OEMs achieve Level 2 and Level 3 ADAS capabilities at competitive prices.

Today, one of the persistent problems associated with ADAS solutions in the industry is the challenge of collecting and leveraging clean and high-quality data. ADAS software is heavily dependent on deep-learning sets and consolidating all that data physically is an industry-wide problem. Inadequate training data affect the solution’s capability to predict unexpected situations which drivers could potentially face on the road.

To improve the solution’s performance and enable superior situational analysis, StradVision supports its widespread deployment on as many vehicles as possible to gather real-time data for training. The firm also provides re-training tools, allowing its clients to train the solution with their own data.

A case in point: StradVision catered to a Tier 1 company that was seeking an ADAS system to leverage the existing cameras and enable AEB (Automatic Emergency Braking) in their vehicles. The company used four surround view monitoring (SVM) cameras in their design to gather visual information and consolidate that into a single bird’s eye view of the car. Fulfilling their need to enable AEB in the same vehicle without the additional implementation of front and rear cameras, StradVision’s software allowed the customer’s cars to achieve flawless AEB using the same four SVM cameras.

StradVision recognizes the demand for self-driving cars in the future and focuses on rapid developments to achieve higher levels of ADAS capabilities. The firm’s development team is improving the solution to help users switch between manual and assisted driving, while ensuring that the driver is ready to take over. A driver monitoring system is another addition to the software, enabling an AI to assess the driver’s alertness and make decisions based on a situation’s threat level.

StradVision has a prominent footprint in China and Germany, with various customers availing its industry-renowned solution. The company plans to expand into the Japanese and Indian markets with pre-production and production programs, while broadening its impact in the US market. StradVision's aim to provide a fully autonomous driving system is opening doors for the adoption of bleeding-edge technologies that are constantly improving people’s lifestyles and road safety.

"Our solution leverages DNN and AI to allow real-time detection of objects, pedestrians and other vehicles"

- Junhwan Kim, CEO

follow on linkedin follow on twitter Copyright © 2019 www.autotechoutlook.com All Rights Reserved | Privacy Policy
Top