Depending on the application, their technology now works with or without the sensor handle that was part of previous versions. The slightest touch on a hand truck or stroller and the minimal variations in the measured driving data are enough for the system to assist—almost intuitively—with acceleration, braking, and turning. The electric motors located inside the wheels provide all the necessary information, without the need for additional sensors.
A parent pushing a stroller or a warehouse worker maneuvering a hand truck: both guide their vehicle by making small changes in the pushing, pulling, or pressing forces they apply. Each of these subtle changes, however small, tells the smart wheels what the user wants and where they intend to go. The wheels then interpret these incoming signals and respond accordingly. Vehicles equipped with this technology are easier to start, can glide more smoothly around corners, and stop more quickly. “With our smart wheels, it’s much easier to move the load because the electric motors in the wheels provide much of the acceleration,” explains Professor Nienhaus. Unruly shopping carts or bulky hospital beds can now be guided with gentle touches, instead of the pushes and pulls normally needed to keep them straight. Bicycle trailers are no longer limited to being passively pulled, and rollators, walkers, and wheelchairs can now actively assist their users.
Nienhaus and his team at Saarland University have researched, developed, and refined this technology over many years. In 2018, the team demonstrated the self-sensing capabilities of these wheels at the Hannover Messe technology fair. They showed that by interpreting measurement data from electric motors located inside the wheels, the wheel system could determine when and how to assist movement, without the need for additional sensors. In 2023, they demonstrated that by introducing a sensor handle, the wheels can move loads with ease: accelerating or braking, turning slower or faster as needed, with each wheel acting independently or in sync with the others. “By combining the sensor handle with our smart wheel technology, we are now able to move more than four tons,” says Matthias Nienhaus. Until now, the sensor handle was necessary as a user interface that allowed the user to specify the direction of movement; it also housed the system's control center: the microcontroller.
Wheels that think for themselves: assisted movement without the sensor handle
In the latest version of their smart wheel technology, Nienhaus and his team have found a way to completely eliminate the sensor handle. “This makes the technology simpler, more intuitive, and more cost-effective,” says Nienhaus. A small change in the pushing or pulling force applied by the user is enough for the wheel drives to recognize whether they should provide gentle support or a more powerful push. This is possible because, with the right motion control strategy, the drives are essentially self-regulating and can generate the necessary control commands themselves. “Precise vehicle control is achieved through a combination of AI methods, high-performance algorithms, and the identification of relevant parameters from the data of the electric motor in the wheels,” explains Nienhaus.
The new handlebar-free technology has been shown to work on vehicles up to 100 kilograms, depending on the specific application and payload. At this year’s Hanover Fair, the researchers will demonstrate how smoothly a bicycle trailer moves when equipped with smart wheels. The team continues to investigate how to transport significantly heavier loads without a sensor-equipped handlebar. For very heavy vehicles, smart wheels combined with a sensor-equipped handlebar remain the best solution.
Artificial intelligence within the wheel: data instead of additional sensors
The electric motors inside the wheels also function as sensors, effectively acting as the wheels' own sensory organs. When a wheel moves, the readings from the drive motor inside it change. The minute variations in these values accurately reveal what the wheel is experiencing. "The electric motors in the wheels provide all the measurement data we need to control them. Therefore, there is no additional cost for installing extra sensors," says Matthias Nienhaus. In several previous research projects, his group studied how to extract as much information as possible from the drives by analyzing the measurement data generated as the wheels rotate. A dataset, for example, could show how the electromagnetic field is distributed at specific points on the motor and how this distribution changes as the wheel turns.
The researchers collected countless such measurements from the drive systems and correlated them with specific motor states and wheel positions. From the vast amount of data collected, they were able to identify signal patterns that correspond to typical mechanical processes. Nienhaus and his team have also filed a patent application for a method that makes motor data even more meaningful by eliminating any interference effects. The measurement data can reveal, for example, the force imparted by a drive, how the wheel positions change, or whether the wheels are more heavily loaded on one side than the other.
Recording small changes in the applied forces is sufficient to keep the wheels on their trajectory.
Mathematical models and intelligent algorithms are used to precisely control the motor drives so they respond appropriately and to monitor their proper functioning. The wheel motors supply the data, and a microcontroller processes the signals. The wheels are extremely sensitive and continuously "detect" their position, speed, and load. Even the slightest movement of the vehicle on which they are mounted—caused, for example, by a touch of the hand—is enough for the drive system to deduce the intended action. The microcontroller reads these signals from the motors, compares them, and provides the appropriate level of thrust. All the information is transmitted via the standard wiring that is already installed.
The research was partly conducted within the framework of cooperative projects funded by the German Federal Ministry for Research, Technology and Space (BMFTR) and the Central Innovation Programme for SMEs (ZIM).
Professor Matthias Nienhaus is keen to see the results of his team's applied research translated into commercial and industrial applications. To that end, he founded WELLGO Systems GmbH as a spin-off company from his university research department.
