Some cool appliations of data science in mechanical engineering

What follows is an incomplete list of cool things that one can do combining mechanical engineering with data science. Note that these applications are rather advanced. We will not learn how to carry them out in this class. You will, however, learn the fundamentals upon which the data science components of these applications are based.

Space X Falcon rocket landing

Space X uses a particular data science technique called Kalman filter  to estimate the position and velocity of the Falcon rocket using noisy data coming from GPS and accelerometers. The characteristic of this technique is that it quantifies the uncertainty in the estimate. Once the estimate is available, then we can decide which thrusters to activate to control the rocket as desired.

Boston Dynamics spot autonomous navigation

Spot is a dog robot. Spot also uses the Kalman filter to localize itself in space. Here is the description of the video: “Spot autonomously navigates a specified route through an office and lab facility. Before the test, the robot is manually driven through the space so it can build a map of the space using visual data from cameras mounted on the front, back and sides of the robot. During the autonomous run, Spot uses data from the cameras to localize itself in the map and to detect and avoid obstacles. Once the operator presses ‘GO’ at the beginning of the video, the robot is on its own. Total walk time for this route is just over 6 minutes.”

Smart extraterrestrial habitats

The Resilient Extraterrestrial Habitats Institute (RETHi) is a NASA-funded project based at Purdue the vision of which is to develop technologies that enable the design of habitats on the Moon and Mars. These structures will likely remain without crew for significant periods of time ranging from months to years making autonomy a key requirement. Prof. Bilionis is leading the Awareness Thrust of the institute. The Awareness Thrust is responsible for developing data science technologies for assessing the health state of all habitat systems using sensor data and controlling robotic agents to carry out maintenance and repair activities. Note that every year there are various opportunities for undergraduate research on data science related topics at the institute. In the video you will see Prof. Dyke, who is leading the project.

Virtual surgery

Prof. Buganza combines physical models of skin with data collected from patients and experiments to simulate skin surgeries. In the process, he has to solve parameter calibration problems, quantify uncertainties in skin properties, and propagate these uncertainties through the physical models.