The self-driving car is one of the most high-profile real world implementations of AI. This is in part because it’s easy to see how such vehicles can revolutionize things like commuting, package delivery, long-distance trucking and so on. While it’s a bit harder to imagine the practical applications of a self-driving motorcycle, Yamaha has gone ahead and built a couple anyway.
A creation of Yamaha’s new Innovation Center, MOTOROiD will make its international debut this week at the Consumer Electronics Show (CES) in Las Vegas.
MOTOROiD’s computer vision and face recognition system allows the user to summon the bike with the wave of a hand. Stability is achieved via an AI-trained “Active Mass Center Control System” (AMCES) which shifts the centre of gravity in real time by rotating parts of the machine such as the battery, swingarm and rear wheel around an axis running through the center of the bike. An inertial measurement system uses a gyro sensor to detect axis rotation and an accelerometer to detect velocity, and sends data to the bike’s onboard control unit with a lag of less than 0.0005 seconds. MOTOROiD has a lithium ion battery, 3D-printed tires, and a haptic HMI device that cradles the rider’s waist. Yamaha describes the riding experience as one in which “the rider resonates harmoniously with the machine.”
At present MOTOROiD only operates at low speeds, but there are plans to push this performance metric in the future.
Yamaha’s other new autonomous bike project is not actually a bike, but a humanoid robot called MOTOBOT which can drive a stock motorcycle without any major modifications. MOTOBOT was developed in collaboration with San Francisco research institute SRI International’s robotics program. Says a Yamaha CES Media Spokesperson, “Yamaha was mainly involved in designing the algorithms for high speed motorcycle riding, analysis of vehicle dynamics using simulations, and reliability design of electronics; while SRI was involved in developing the robot based on our requirements, and developing the position sensing system.”
MOTOBOT uses its actuators to manipulate a stock bike’s steering, throttle, brake, clutch and shift controls in much the same way a human would. Unlike its relatively sluggish cousin MOTOROiD, MOTOBOT is a speed freak that has already surpassed its project’s straight line velocity target of 200 kph.
One of the MOTOBOT development advisors is Italian motorcycle racing champion Valentino Rossi. Yamaha’s goal is to get its bot up to the level where it can better Rossi on a racetrack.
Last October at the Thunderhill Raceway Park in Sacramento Valley, California, MOTOBOT clocked a lap time of 117.5 seconds compared to Rossi’s 85.7 seconds on the two mile (3.2 km) circuit. SRI says refinement of the bot’s high-precision GPS, sensor fusion and machine learning technologies will enable MOTOBOT to improve its decisions and optimize track line and race performance. Yamaha boldly predicts the bot will outperform Rossi within two years. Is that possible?
It’s worth noting that skeptical AI researchers had estimated no machine would beat a top human in the ancient board game Go for another ten years — that is, until DeepMind’s AlphaGo dispatched world-best Ke Jie last spring in Beijing. After the series, professional Go players began adopting many of AlphaGo’s novel opening strategies into their own games. Might MOTOBOT one day similarly introduce humans to new and more efficient ways of racing a motorcycle around a track?
Only time will tell. However, like AlphaGo, MOTOBOT is more a proof of concept project. Yamaha says one of the possible practical applications might be using MOTOBOT as a test driver. “We hope to visualize the human side of motorbike riding and use the information gathered to identify the relationship between the rider and the bike. That would help us to develop more exciting vehicles in future. We could also adopt the control and maneuvering programs to other vehicles such as marine jets and snowmobiles.”
It’s clear that even traditional manufacturers such as Yamaha cannot afford to ignore AI’s potential for industry disruption, and so are driven to embrace the tech, using it to reimagine their products and push limits. With BMW and Honda also working on self-driving bikes, Yamaha’s first-ever appearance at the CES is bound to rev up this two-wheeled innovation race.
Journalist: Michael Sarazen