Innoviz Technologies Lidar-Based Perception Stack for Global Automotive OEM

Innoviz Technologies Ltd announced on August 4 that it is developing a lidar-based perception stack in collaboration with a top 10 global automotive OEM for future vehicle programs. The project combines the company's InnovizTwo lidar technology with advanced perception software running on NVIDIA hardware. The jointly developed solution will be designed and validated for a highway operational design domain (ODD), allowing the automotive manufacturer to evaluate the technology for potential integration into upcoming vehicle platforms.

Highway Perception Stack Combines Lidar and AI Processing

The new perception stack is built around Innoviz's lidar technology and software capabilities to support advanced highway driving applications. It incorporates core Physical AI functions that transform lidar data into an understanding of the surrounding environment. By combining InnovizTwo sensors with NVIDIA computing hardware, the system is intended to deliver reliable perception performance for future automotive programs while supporting validation activities required for highway operational design domains.

Key Functions of the Lidar-Based Perception Stack

The perception platform includes multiple capabilities designed to improve environmental awareness and hazard detection for automated driving applications.

• Point Cloud Plus (PC+) processing for enhanced scene interpretation.
• Obstacle Detection that identifies drivable and non-drivable areas.
• Detection of in-path hazards including tires, lost cargo, and Euro pallets.
• Recognition of road users such as pedestrians and vehicles.
• Classification of detected objects into pedestrians, cars, trucks, and two-wheelers.

Hazard Recognition Supports Future Highway Programs

The perception stack classifies every pixel within the driving scene as either drivable space or a non-drivable obstacle, enabling the system to identify potential hazards directly in the vehicle's path. It can detect objects such as tires, lost cargo, and Euro pallets while simultaneously recognizing surrounding road users including pedestrians and vehicles. The software also categorizes detected objects into classes such as pedestrians, passenger cars, trucks, and two-wheelers, providing the detailed environmental understanding required for future highway-focused automated driving systems.

Frequently Asked Questions

What did Innoviz Technologies announce regarding its latest lidar project?
Innoviz Technologies announced the development of a lidar-based perception stack with a top 10 global automotive OEM for future highway driving programs. The solution combines InnovizTwo lidar technology, advanced perception software, and NVIDIA hardware while supporting a highway operational design domain. The system includes Point Cloud Plus processing, obstacle detection, hazard identification, and object classification to help evaluate its suitability for future automotive applications.