I've been involved in developing Honda robots since 1986, working with a team of people who feel strongly that the most important aspect of a robot is its ability to help people when they are in need. Our development work focuses on enabling a task-performing robot arm utilizing ASIMO technology to help in the country faces difficulties of the March 2011 accident at the Fukushima Daiichi Nuclear Power Plant. Additionally, our walking assist devices are being used to help elderly evacuees at temporary housing in the aftermath of the Great East Japan Earthquake.

We were engaged in research for the new ASIMO at the time of the earthquake and made a proposal to the president based on our desire to put our robot technology to use restoration from the accident. Despite the fact that development is an expensive undertaking, the president immediately approved our request. Just three months after the earthquake, we had begun to develop the task-performing robot arm after speaking to TEPCO officials about what our technology could do. The most challenging part of the project was determining the robot's specifications since it was unclear what was going on at the accident site. We didn't know how much radiation the device would need to be able to withstand, or to what other harsh conditions it would be exposed.

Later, it became clear that crews needed a robot that could reach places deep inside the rubble-strewn facility, and we realized that the technologies used by ASIMO could play a useful role. For a robot to open and close valves in the facility, the hardware would need to exhibit both autonomy and compliance (flexibility and elasticity). For example, just as it is difficult to change a light bulb on the ceiling if there's no stable place to stand, the robot needs technology capable of providing sure support even when its base lacks stability. We addressed this challenge by using exactly the same control technology that we used to ensure ASIMO could keep its footing, enabling the task-performing robot arm to complete tasks while forecasting and estimating the angle of the valve handle and type of floor surface.

In this robot development project, it was not clear what the site would require of our hardware. Yet ultimately, we were able to provide a solution by setting our sights on creating a robot capable of completing the same repair and maintenance actions in the facility as a person could accomplish directly with his or her hands.