Raytheon to Utilize Akida Neuromorphic AI Chip for Radar Processing

Raytheon has partnered with Brainchip to utilize the Akida neuromorphic AI chip in a US research project focused on radar signal processing. The $1.8 million contract from the Air Force Research Laboratory (AFRL) was announced in December of last year. The project aims to explore the mapping of complex sensor signal processing algorithms onto neuromorphic chips. Raytheon, a part of the RTX conglomerate, has been selected to provide services and support for the contract, making it a crucial collaboration for Brainchip.

The Akida neuromorphic AI processor is specifically designed to handle neural networks and machine learning algorithms with ultra-low power consumption. Its spiking networks can be effectively mapped to radar signals for edge processing systems, showcasing its versatility and efficiency in processing tasks.

• Brainchip launches edge AI box ecosystem
• Akida AI chip for epilepsy smart glasses
• Akida spiking processor could heat for FD-SOI process

Sean Hehir, the CEO of BrainChip, emphasized the importance of minimizing system size, weight, and power (SWAP-C) for radar signaling processing on ever-smaller mobile platforms. He highlighted the improved radar signaling performance per watt achieved through neuromorphic computing, demonstrating significant benefits for mission-critical applications. This collaboration with Raytheon underlines the potential of neuromorphic chips in enhancing radar signal processing capabilities.

Hehir further explained that radar signaling processing is crucial for various airborne and mobile platforms, necessitating the optimization of system size, weight, and power. The contract with the Air Force Research Laboratory signifies a milestone in enhancing radar signaling applications through low-power, high-performance computing. The recognition from leading organizations like AFRL validates the effectiveness of Akida hardware and state-space AI models utilizing Temporal Enabled Neural Network (TENNs) model offerings.