The auto giant said last week that it successfully completed an initial test in which one of the company’s experimental robots helped disabled war veteran Romulo Camargo in Florida with basic tasks in his home for two days.

Romulo Camargo, who became paralyzed from the neck down after surviving a gunshot to the back of his neck in Afghanistan, used Toyota’s so-called human-support robot in a two-day trial last November, said Doug Moore, Toyota’s senior manager of future mobility business, technology for human support.

Get Data Sheet, Fortune’s technology newsletter.

The goal of the trial was to learn how best to use robots as helpers to the disabled and act as care-takers. It’s part of a Toyota research initiative exploring robotics and the related potential business opportunities, Moore explained.

Toyota debuted its human-support robot in 2012 and has been testing it in Japan. The white robot, which can weigh up to 80 pounds, resembles an oversized microscope outfitted with an arm that can grip onto objects.

The car company had a pre-existing relationship with Camargo and first discussed with him and his family what roles a robot could play in the household. The parties outlined about 20 different tasks that the robot could help with, but ended up choosing just three.

The first task the robot was able to help with involved helping Camargo get into his house in case he accidentally locked himself outside.

Camargo’s home contains a special button installed inside near the front door that he can press to let him out of the house. However, if Camargo leaves the house and the door accidentally shuts behind, he can’t go back inside because there is no button on the outside of the house that he can press to re-open the door.

To rectify the dilemma, Moore explained how Toyota was able to connect the robot to an outside wireless security camera. Because the robot was programmed to recognize Camargo’s face, the robot could tell that Camargo was stuck outside the home via the wireless security camera. The robot was then able to move itself to the button near the doorway, and used its arm to press the button that opened the door, which let Camargo back inside.

The other two tasks involved programming the robot to bring Camargo both snacks on a food tray, and water in a bottle. Similar to the way warehouse robots can roam through facilities on their own, Toyota’s robot was able to move autonomously without bumping into objects in Camargo’s home using an internal map of the house, Moore explained.

Toyota decorated objects in the house, like the water bottle or chairs, with QR codes so it could locate the respective objects. Although advances in computer vision technology has made it possible for computers to recognize objects better than they have been in the past, it’s still not economically viable for Toyota to use all the various vision technologies, like Google’s (GOOG) cloud vision APIs, in its robots, Moore explained.

Eventually, Toyota hopes to incorporate into its robots specialized chips that can aid with specific computer-vision tasks in addition to computer vision software sold via cloud computing services to make more powerful robots that can “see” the world.

The challenge is making a powerful robot that’s relatively affordable, which as Moore explained, is not yet ready in today’s market. Using Toyota’s human-assisted robot to care for the disabled is still very much a research project, but the hope is that one day, it can become commonplace.