stradvision

SEOUL, South Korea - April 7, 2025 - STRADVISION, a global leader in AI-powered vision perception technology, today unveiled a three-year product roadmap aimed at addressing the evolving autonomous driving landscape and meeting customer demands. At the heart of this roadmap is SVNet MultiVision Gen 2, a next-generation vision perception solution slated for completion by 2027, with a prototype debut planned for CES 2026. 

Advancing Towards L3/L4 Autonomous Driving with a Two-Track Strategy 

STRADVISION is expanding beyond L2/L2+ Advanced Driver Assistance Systems (ADAS) to develop L3/L4 autonomous driving solutions tailored to industries including robotaxis, logistics, public transportation, and agriculture, through its Two-Track strategy. 

Unlike L2 systems, where the driver remains responsible for vehicle control, L3 and above require the system to assume full driving responsibility. To achieve this, STRADVISION is focusing on end-to-end (E2E) model development, integrating planning, path prediction, and driving control alongside perception—a key driver of this new roadmap which is expanding from individual modular-based models to end-to-end (E2E) model development.  

SVNet MultiVision Gen 2: Enhanced Camera Coverage, HD-Mapless Functionality, and Cutting-Edge AI 

SVNet MultiVision Gen 2 substantially enhances its vision perception capabilities by supporting 11 camera channels—six more than its predecessor, which combined FrontVision and SurroundVision. This surpasses the number of cameras found in many currently available L3 autonomous vehicles, enabling more precise environmental recognition. Additionally, it supports HD-Mapless autonomous driving (ADS) L3/4 functionality, addressing limitations associated with Geo-fencing-based navigation. 

Designed as a scalable and flexible solution, MultiVision Gen 2 will offer seamless integration across various camera systems and autonomous driving levels, ensuring compatibility with a wide range of customer product lines. 

STRADVISION will also integrate cutting-edge AI technologies, including Vision Transformer and Vision Language Models (VLMs), to optimize performance in diverse driving environments. By building on its existing vision perception expertise, the company aims to develop a full-stack E2E solution for L3/L4, applying a modular approach to enhance adaptability and efficiency. 

STRADVISION CEO Junhwan Kim emphasized the company’s commitment to advancing autonomous driving technology, stating: “By 2027, we will collaborate closely with leading SoC (System-on-Chip) partners to develop MultiVision Gen 2 into an embedded solution optimized for mass production across various SoC platforms. Through this innovation, we aim to enhance road safety and move closer to STRADVISION’s founding mission: protecting the lives of drivers and pedestrians.” 

Shaping the Future of Autonomous Driving 

With this three-year roadmap, STRADVISION continues to push the boundaries of autonomous driving technology, solidifying its competitive edge and accelerating the commercialization of L3/L4 autonomous vehicles. 

For more information about the groundbreaking technology, SVNet, please visit STRADVISION.