Abstract:

At Roanoke College, our introductory programming course uses project-based learning with multiple themes, robotics, video games, and data mining, to engage students and connect their learning to real-world applications. Teaching the robotics version of the course in the time of COVID-19 is particularly difficult because of the personal robot hardware. In our course, students form groups to collaborate on programming projects using the robots throughout the semester. Besides reducing cost, sharing robots helps establish peer groups that can improve student learning and attitudes. However, teaching during a global pandemic means that the course must support students learning remotely. Teaching the class with physical robots would preclude students from sharing hardware and would make collaboration more difficult.

We decided that the best way to support remote collaboration between students on programming activities was to teach the course using a robotics simulator. Simplistic robotic simulators have been used for decades to help students understand a notional machine and have a well-established efficacy. However, these systems abstract robotics to the degree that they do not satisfy the course theme's intent to connect students to real-world applications. There are also sophisticated robotics simulators that accurately mimic actual hardware. However, these simulators require students to install software, which burdens students to solve installation and compatibility issues.

To make learning remotely as easy as possible for our students, we developed a web-based simulation that uses a 3D physics engine to visualize program execution. We designed the simulator to be realistic but not to emulate any existing robot hardware. As a result, it feels authentic to programming a physical robot. However, because it is not restricted to being accurate, users can create environments that would be impossible in the real world. We tested the simulator in our introductory computer science class. We hoped that the simulator would allow our students to achieve the same learning outcomes while learning remotely, despite the simulator being less engaging than physical robots.

What we found instead was a plethora of benefits. The simulator was less expensive, less work for us to maintain, and more manageable for students to develop. Additionally, students could tweak the physics simulation parameters, and they had no fear of damaging expensive equipment. These differences encouraged students to play and experiment with the simulation. They could accelerate their robots to incredible speeds, jump them off of ramps, crash them into destructible walls, and allow them to plummet into limitless depths. The simulator also allowed students to easily create robots and interactive environments by positioning and constraining geometric primitives. Students had fun designing challenges and sharing their creations with their peers, and while using the simulation, students engaged with programming activities more than in previous iterations of our robotics course.

Keywords:

Remote learning, personal robots, simulator, engagement.