Visual Servoing Using Edge AI

0001-01-01

Artificial Intelligence (AI) is transforming many facets of life. In robotics, there is a noticeable gap between the digital and physical worlds, making it challenging to deploy AI-enabled robots in dynamic settings such as factories or warehouses. Sending data from a robot to the cloud degrades performance by introducing significant latency and jitter, especially in vision-based applications given the large size and high bandwidth requirements for raw image data.

Edge AI eliminates the need to offload raw data, greatly improving performance. Analog Devices’ MAX78000 AI microcontroller enables AI at the edge with its convolutional neural network (CNN) accelerator. The CNN accelerator executes AI inferences in hardware, resulting in faster insights versus those performed in software on other microcontrollers. In robotic applications, local AI dramatically reduces latency in tasks such as visual servoing, increasing performance. In addition, the MAX78000 boasts one of the industry’s lowest power consumptions. AI inferences can be performed at the edge using a just coin-cell battery, enabling true edge intelligence in mobilized battery-powered applications.

Visual Servoing Using Edge AI

0001-01-01

Artificial Intelligence (AI) is transforming many facets of life. In robotics, there is a noticeable gap between the digital and physical worlds, making it challenging to deploy AI-enabled robots in dynamic settings such as factories or warehouses. Sending data from a robot to the cloud degrades performance by introducing significant latency and jitter, especially in vision-based applications given the large size and high bandwidth requirements for raw image data.

Edge AI eliminates the need to offload raw data, greatly improving performance. Analog Devices’ MAX78000 AI microcontroller enables AI at the edge with its convolutional neural network (CNN) accelerator. The CNN accelerator executes AI inferences in hardware, resulting in faster insights versus those performed in software on other microcontrollers. In robotic applications, local AI dramatically reduces latency in tasks such as visual servoing, increasing performance. In addition, the MAX78000 boasts one of the industry’s lowest power consumptions. AI inferences can be performed at the edge using a just coin-cell battery, enabling true edge intelligence in mobilized battery-powered applications.