In this final recipe, we are going to use SenseHat on the Raspberry Pi to collect data, train that data on our local computer, then deploy a machine learning model on the device. To avoid redundancy after recording your data you will need to run either of the recipes on autoencoders or isolated forest from earlier in this chapter.
People use motion sensors in IoT to ensure shipping containers are safely transported aboard ships. For example, proving that a shipping container was dropped in a particular harbor would help with insurance claims. They are also used for worker safety to detect falls or workers acting unsafely. They are also used on devices that are prone to vibration when malfunctioning. Some examples of this are washing machines, wind turbines, and cement mixers.
During the data collection phase, you will need to safely simulate falling or working unsafely. You could also put a sensor on a washing machine that is unbalanced....
