Apera Innovates Intelligence for Robotics, Bringing ‘4D Vision’ to Automated Manufacturing

A Vancouver company advancing intelligence in robotics is competing for “Company of the Year – Growth” at this year’s Technology Impact Awards powered by the BC Tech Association.

Apera AI was founded in 2017 by Sina Afrooze and Armin Khatoonabadi. Afrooze, who serves his company as chief executive officer, is an expert in the deep learning space, with previous stints at Avigilon and Amazon Web Services.

His firm’s flagship technology is “4D vision,” which brings “human-like sight” to automated manufacturing processes.

Built around proprietary artificial intelligence and machine learning technologies that control object recognition and robotic pose estimation and path planning, 4D Vision is run through robotic workcells using Apera Vue vision software.

Vue is robot-agnostic and compatible with many leading robot brands, according to Afrooze, and gives robots new powers such as significant faster decision-making abilities. This provides customers with greater certainty that there will be a return on their investments into automation.

The process begins with a 3D scan. The part is trained into a digital twin environment using synthetic data, and only when the training model reaches a reliability level of 99.99% does Apera supply a program to the customer. Training up to this level of reliability often takes up to one million cycles.

The use of synthetic data here is intentional, according to the founders. In training its neural networks with synthetic data, Apera can bypass the need for reference images as well as reduce reliance on CAD models.

“Synthetic data allows us to improve the AI process and introduce variables such as lighting, texture, and labels or markings,” the company says online.

Currently, the company is targeting key sectors, including automotive manufacturing, metal fabrication, and the medical devices field.

Afrooze, who graduated from Simon Fraser University with a Bachelor in Systems Engineering, holds 10 patents across the fields of AI, digital imaging, image processing, and media streaming.