The End-To-End Quickai Hardware / Software (HW/SW) Platform Can Be Implemented Quickly and Easily with Minimal Data Science and Firmware Resources
QuickLogic Corporation, a developer of ultra-low power multi-core voice-enabled SoCs, embedded FPGA IP, display bridge, programmable logic and Endpoint AI solutions, announced that mtes Neural Networks (mtesNN) of Japan has selected its QuickAI™ Platform for a new generation of AI-enabled endpoint devices. With the complete end-to-end QuickAI Platform solution, mtesNN is accelerating new product designs that leverage the benefits of local decision making based on real-time sensor data. This significantly lowers decision latency, improves reliability and eliminates the cost and high power consumption of a full-time broadband connection to cloud-based AI processing.
mtesNN was founded in 2015 and has developed expertise in structural health and surveillance monitoring with sensor modules and cognitive cameras that leverage the benefits of AI. Sensor modules will be deployed around cities to analyze the impact of earthquake tremors on railways, bridges and tall buildings so that preventive maintenance can be deployed. Cognitive camera systems will be deployed for surveillance and actionable event detection along streets and in large venues to improve safety and provide real-time information for emergency responders.
The challenge mtesNN faced was that cloud-based AI processing is simply not suitable for its endpoint applications. Cloud-based AI systems require broadband connections to send large amounts of raw sensor data and images to the cloud for processing and decision making. This requirement increases system costs, operating costs, latency, power consumption and the risk of downtime.
With the QuickAI Platform, the inferencing (decision making) is done locally with substantially reduced latency and power consumption. This is particularly important for mtesNN, which depends on solar power with battery back-up in some applications. The QuickAI Platform also allows mtesNN to lower system and operating costs and improve reliability by eliminating the need for continuous high-bandwidth connectivity. Because the QuickAI hardware and software Platform enables AI endpoint solutions to be developed easily and quickly with minimal data science and firmware engineering resources, mtesNN is also benefitting from lower product development costs while gaining valuable time-to-market advantages.
“We evaluated numerous design approaches before selecting QuickLogic’s QuickAI Platform to develop new AI-enabled endpoint devices that leverage the many benefits of local AI processing,” said Takaro Harada, CEO of mtes Neural Networks. “With QuickAI’s end-to-end hardware and software, we are able to extend battery life while accelerating our new product development cycles. We are excited to use the QuickAI Platform for this and future generations of AI-enabled endpoint devices.”
“We are very happy that mtes Neural Networks chose our QuickAI Platform to enable endpoint artificial intelligence in their next generation of AI-enabled endpoint devices,” said Brian Faith, CEO of QuickLogic. “The unique heterogeneous multi-core architecture and end-to-end hardware /software solution provided by QuickAI Platform simplifies and accelerates the implementation of AI by providing standard interfaces to sensors and traditional digital computing resources while leveraging leading edge Neural Processing (NPU) technology. We look forward to continuing our work with mtes Neural Networks as it develops new AI-enabled endpoint devices.”