AI Jackal roams UT-Austin.

Like many young humans, Jackal is learning to navigate the world. The rover is part of an autonomous robotics project at the University of Texas, where students like third-year computer science student, Luisa Mao, are teaching it to navigate outdoor terrain. It's a challenge; Jackal has run off, crashed, and rammed into people. "I think of it as a pet that doesn’t do what it’s trained to do all the time," Mao said.

“I think of it as a pet that doesn’t do what it’s trained to do all the time.” ― Luisa Mao

A Challenging Task

Mao's lab, AMRL, is funded by industry and government sponsors, and specializes in AI robots that operate independently. These robots could have a range of applications, from food delivery to search and rescue. The challenge? Teaching them to navigate unknown environments, a problem Mao predicts will take years to solve.

“We’re looking at unknown environments, so it’s a more difficult problem to solve. It’s going to take a long time to develop, years probably, both from the perspective of creating the technology and adapting cities to robots.” ― Luisa Mao

How Does It Work?

Central to the project is teaching Jackal to understand its surroundings. Autonomous vehicles like self-driving cars use lidar to detect obstacles, but they can't interpret the world visually. Jackal uses a camera and AI to make decisions about where to go, identifying different terrains and choosing the best route. Its "brain" is a neural network of nodes, taking up just 2GB of storage. Mao and her team can see what the robot sees: a view of the ground, with potential paths charted. Jackal's task? To navigate to the end of a courtyard using the sidewalk. "We switched to autonomous mode," said Zhiyun Deng, a doctoral student. And off Jackal went.

Sources/URLs:

• [How a University of Texas robot named Jackal is learning to navigate the world](https://www.statesman.com/story/news/education/2023/02/19/how-university-texas-robot- named-jackal-learning-navigate-world/6929734001/)