Motion sensor learns mistakes in fitness routines
At next month’s CES, Bosch Sensortec will unveil a sensor that is one step closer to a personalised trainer. The BHI260AP revolutionises fitness tracking, says the company by recognising exercises based on users’ different movements, and styles of movement, as well as the equipment used, body sizes and performance levels.
The self-learning motion sensor uses artificial intelligence to detect and respond ot a wide variety of movements. It can understand when the wearer is doing squats, sit-ups, push-ups or kettlebell training, for example. All of these require different body positions and movements depending on the equipment used.
The BHI260AP in wearable and hearable devices automatically detects and responds to movements and pre-learned fitness activities. It is able to learn any new fitness activity based on regularly repeating sequences and cyclical patterns.
Existing activities can be personalised and adapted to the user’s own individual style, adds Bosch Sensortec. Devices using BHI260AP sensors automatically track fitness activities and analyse the training intensity, which means that users can be trainers and trainees at the same time.
The BHI260AP is suitable for use in always-on sensor applications such as fitness tracking, pedestrian positioning, machine learning analytics and orientation estimation. The smart sensor includes a variety of software functionalities, including the self-learning AI software, a 32-bit customer programmable microcontroller, and a six-axis IMU (inertial measurement unit) all in one package.
Typical applications are personalised fitness tracking, swim analytics and pedestrian always-on position tracking.