Designing systems that help people and society through model-based control

At Keio University, the Takahashi Laboratory, in the Department of System Design Engineering (Faculty of Science and Technology) is doing research to develop products and systems that are useful to society, through an approach called model-based control.


In model-based control, the controlled system, such as a spacecraft, automobile, or robot, is modeled mathematically, and control technology is developed for the model. While this method has the advantage of enabling control technology to be fine-tuned while doing simulation, the technology may not work as planned with the simulation if the precision of the model is low, or the operational environment differs from the actual one. For this reason, it's important to achieve accurate modeling that anticipates the operational environment, and highly robust design that can respond to changes in the environment.


Q "In our lab, we handle a very diverse range of systems, including spacecraft and aircraft. We also handle ordinary means of mobility, such as automobiles and wheelchairs, as well as vibration control in buildings. The range is so diverse, at first glance, our research might seem a "mixed bag." But because moving objects are based on certain physical phenomena, if we can describe those physical phenomena mathematically, we can build models. And if we apply control theory based on those models, we can obtain the desired control performance."


This is a walking measurement robot, developed to prevent elderly people from falling. It uses a laser rangefinder to measure through a range of 240 degrees and 4 meters. In this way, it estimates the position and velocity of both legs, while detecting and tracking the characteristic shape of a leg. From that information, this system enables walking to be measured while guiding the user, by building in control technology to maintain a specific distance and velocity.


To achieve high-precision model-based control, it's essential for modeling to include the operational environment. An especially difficult aspect is anticipating how users will act. To achieve that, it's necessary to predict and evaluate human behavior, and respond appropriately.


In research on controlling inverted-pendulum-type mobility systems, such as the Segway, the Takahashi Lab conducts verification tests by people using actual machines. The aim is to propose human models and to analyze how control systems affect the behavior of people, in order to suggest safe, comfortable means of mobility.


Q "Our aim is to be helpful to people and society. So, the first thing we do is field research, to find out what technologies are actually needed. For example, when we take robots into hospitals, to develop robots that can carry medicines, various requirements emerge. We're moving ahead with our research in ways that could resolve each of those issues while utilizing control technology."


Q "For example, regarding wheelchair research, quite a lot of people can't use a joystick. As the population ages, there'll be a rapid increase in wheelchair users, so new user interfaces will appear. Accordingly, specialists in signal processing are creating various technologies, such as measuring brain signals, voluntary blink signals and so on. But in many cases, they don't have technologies relating to mobility, so I think that's one strength of our laboratory. By teaming up with signal processing specialists, I think we can develop means of mobility, such as automobiles and wheelchairs, with new user interfaces. Regarding the measurement of walking, we'd like to focus on collaboration between healthcare and engineering specialists. When users actually gather to discuss this topic, we don't have much knowledge of how to feed the results back to users. So, we'd like to pursue that aspect of the research by bringing in expertise from people in the Faculty of Medicine."