Control system and method based on sensor network and in-memory computing neural network

Abstract

The invention discloses a control system and method based on a sensor network and an in-memory computing neural network, and the invention relates to the field of artificial intelligence. The invention includes a sensor network, a data self-processing module, an analog-to-digital converter module and an in-memory computing neural network module, combines the in-memory computing technology with the sensor network, and utilizes the sensor network with data self-processing capability to increase data bandwidth and Reduce power consumption, complete the calculation of the neural network through in-memory computing technology, avoid too many unnecessary data transmission operations, while ensuring high-precision calculations, greatly improve the calculation speed, reduce data storage space, and reduce the system cost. Overall computing power consumption.

Description

Technical field [0001] The present invention relates to the field of artificial intelligence, specifically to a control system and method based on a sensor network and an in-memory computing neural network. Background [0002] A sensor network is a distributed array of sensors consisting of a large number of tiny, low-cost, feature-rich, information-acquiring sensors that are connected to each other over a wireless network to communicate. The application prospect of sensor networks is very extensive, and it plays a pivotal role in many fields and practical application scenarios, and has great potential for development. With the continuous advancement of microelectronics-related technologies, sensor network technology is also advancing, and has a huge impact on modern information technology industry, modern manufacturing, and modern military information. However, for most practical application scenarios, in order to obtain more accurate environmental information, the sensor networ...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a control system and method based on sensor networks and memory computing neural networks. In order to solve the problem of running and computing bottlenecks in existing sensor networks and traditional neural network hardware Combined with the computing neural network, a data self-processing module is embedded in the sensor network, which effectively filters the invalid data collected by the sensor, and transfers the unit that originally needs to perform data calculation outside the memory to the memory, which is convenient while storing. The calculation process is completed, thereby effectively avoiding a large number of I / O operations in the traditional computer architecture, saving data processing time, improving the overall calculation efficiency of the system, reducing overall power consumption and saving data required of storage space

Images