DEVELOPMENT OF A VIRTUAL LABORATORY TO TEACH MOBILE ROBOTICS

Nowadays mobile robots have become a very important tool in many situations. Thanks to their flexibility, we can find mobile robots in a wide variety of environments, including in industry. This is why mobile robotics has gained relevance in engineering-related degrees. It is very important that students learn all the necessary concepts to design and implement mobile robots that are able to solve a task autonomously.

Every robot must be equipped with a set of sensors that gives it information from the environment where it moves. Using this information, the robot must be able to estimate its position in this environment and to calculate the necessary control action to arrive to the target points. This way, solving correctly the localization problem is crucial for a robot to be autonomous.

Vision sensors are now one of the most used sensors to solve the localization problem. The cost of the cameras is relatively low, comparing to other sensors, and they offer a lot of information from the environment. However, it is necessary to process the images to extract relevant information which is useful to estimate the position of the robot. Nowadays, there are many approaches to extract this information and they lead to many different localization algorithms which may be conceptually very complex.

Taking into account these facts, we have developed a virtual laboratory to teach mobile robotics and computer vision. This virtual laboratory has been designed to be used by the students of a Master in Robotics. In this master, there is a subject whose main objective is learning techniques for estimating the localization of a mobile robot using images.

We have included in this laboratory several databases of images, captured along different rooms of an indoor environment. With these images, students can test the algorithms they learn at the classroom. Since they work with real images, they will face the usual problems that outcome in real working situations, such as occlusions in the scenes, changes in the lighting conditions, etc.

According to our experience in this kind of subjects, during the practical sessions students are usually more concerned about the code they have to program and they tend to forget the final goal they have to achieve. To overcome it, the laboratory includes most of the localization algorithms taught at this subject. Students can configure easily the parameters of each algorithm and the results are show graphically. Also, students can make comparative studies to decide when each algorithm works better and which is the optimal value of the parameters.

Thanks to this tool students will be able to carry out the practical sessions in a very flexible way. We expect that this laboratory helps students to understand the basic concepts that appear in mobile robots localization and to develop the ability to design autonomous mobile robots than can work in a real environment.