Aeva Adopts Cadence Tensilica Vision DSP for Next-Generation 4D LiDAR

Aeva adopts Cadence Tensilica Vision DSP IP to enhance real-time 4D LiDAR performance, efficiency, and scalability.

Cadence announced that Aeva has licensed Cadence Tensilica Vision DSP IP to accelerate signal processing in its 4D LiDAR systems, helping deliver flexible and scalable solutions for industrial robotics and automotive applications.

The collaboration marks a significant step forward for real-time perception and autonomy, combining Aeva’s next-generation 4D LiDAR technology with Cadence’s low-power, high-performance DSP architecture.

Enhancing Real-Time Perception for Robotics and Automotive

LiDAR technology plays a critical role in robotics, autonomous driving, industrial automation, and edge AI applications by enabling accurate 3D mapping and object detection. Aeva’s 4D LiDAR technology expands these capabilities further by simultaneously detecting both velocity and position, enabling autonomous systems to make safer and more intelligent decisions in real time.

To support these workloads, Cadence Tensilica Vision DSPs provide the programmability, scalability, and efficiency required for advanced LiDAR signal processing pipelines. Their customizable architecture and low-power design are particularly suited for latency-sensitive perception workloads.

Cadence Tensilica Vision DSP Delivers Flexible and Efficient Processing

Cadence noted that the Tensilica Vision DSP architecture enables Aeva to add flexibility and customization across its signal processing stack while maintaining power efficiency for edge deployments.

The DSPs also leverage the Tensilica Instruction Extension (TIE) language, allowing hardware-level optimization for specific workloads. This helps improve processing efficiency for demanding LiDAR applications where responsiveness and low latency are critical.

“Cadence’s DSP technology provides the flexibility and performance uplift we need to push the boundaries of perception and deliver scalable solutions to our industrial and automotive customers,” said James Reuther, chief engineer of Aeva.

According to Aeva, integrating the Tensilica Vision DSP into its next-generation LiDAR systems enables the company to take advantage of Cadence’s configurable hardware and optimized software libraries to accelerate development and deployment.

Optimized Software Stack for AI and Vision Workloads

Cadence’s Tensilica Vision DSP platform includes software libraries optimized for neural networks, computer vision, simultaneous localization and mapping (SLAM), radar processing, and point cloud workloads.

The platform also incorporates the NeuroWeave SDK, Cadence’s third-generation neural network compiler designed to support the latest AI models increasingly used in LiDAR and autonomous perception systems.

“Our Tensilica DSPs empower SoC providers to deliver differentiated performance and power efficiency in leading-edge, low-latency applications like lidar,” said Amol Borkar, group director of product management and marketing for Tensilica DSPs at Cadence.

Cadence Expands Presence in Physical AI and Autonomous Systems

Aeva joins a growing list of companies adopting Cadence Tensilica DSPs for industrial robotics, autonomous systems, and physical AI applications. Cadence continues to position the Tensilica platform as a customizable and power-efficient solution for next-generation edge AI workloads spanning ADAS, radar, audio, and vision processing.