Mobile Robots: Perception & Navigation

Today robots navigate autonomously in office environments as well as outdoors. They show their ability to beside mechanical and electronic barriers in building mobile platforms, perceiving the environment and deciding on how to act in a given situation are crucial problems. In this book we focused on these two areas of mobile robotics, Perception and Navigation. This book gives a wide overview over different navigation techniques describing both navigation techniques dealing with local and control aspects of navigation as well es those handling global navigation aspects of a single robot and even for a group of robots. This chapter investigated the state of the art of SLAM system whose aim is to simultaneously map the environment and provide accurate positioning of the vehicle where the state of both vehicle and landmarks are brought into a single estimation process. The estimation process supplies estimation of vehicle and landmarks in terms of mean and variance-covariance estimate of the state vector conditional on the whole set of measurements. The construction of the map involves appropriate initialization step in which some prior knowledge regarding the environment is reported, which allows us to ensure geometrical validation test later on, and an alignment stage in which the observations are turned into landmark Cartesian coordinates. Next, a data association stage is required to map the observations to the already identified landmark or initiate new landmarks. For this purpose, one uses Mahanalobis distance to match the observation to possible landmarks. If none of the landmarks matches the current observation and both the geometrical and statistical tests were positive, then a new landmark is initiated and the state vector is therefore augmented. Note that, in order to balance the cost effectiveness and optimality requirement, the reasoning is carried out only within a submap of the environment, where it is most likely to find the matching given the sensor limitations. At later stage, the obtained landmarks are also used to extract feature landmark consisting of segments and corners. The methodology has been validated in a platform using two Khepera robots, one of which is equipped with vision turret while both are equipped with range infrared sensors and encoders. A virtual interface showing the robot trajectory as well as environment is developed using OpenGL platform for 3D visualization. The use of both robots also allowed us to test and validate collaboration scenarios in which multi-robot localization technique is used in conjunction with SLAM algorithm. The tests carried out demonstrated the validation of the developed algorithm and the consistency of the outcomes when looking at the 95% confidence bound limits. This open a new area of research where more advanced collaboration scenarios can be used in more complex environments where the features can be constituted of either geometrical or non-geometrical features. On the other hand, inspired by the overall intelligent behaviour of large biological insect communities, together with the rapid development of the field of distributed artificial intelligence, through, for instance, the concrete RoboCup robot soccer initiative, this offers new motivation grounds for further developments of multiple robot systems at different research communities The proposed navigation system, based on a topological representation of the world, allows the robot to robustly navigate in corridor and structured environments. This is a very practical issue in assistance applications, in which robots must perform guidance missions from room to room in environments typically structured in corridors and rooms, such as hospitals or nursing homes. Although the topological map consists of very simple and reduced information about the environment, a set of robust local navigation behaviors (the actions of the model) allow the robot to locally move in corridors, reacting to sensor information and avoiding collisions, without any previous metric information. Another important subject in robot navigation is robustness in dynamic environments. It is demonstrated that topological representations are more robust to dynamic changes of the environment (people, obstacles, doors state, etc.) because they are not modelled in the map. In this case, in which local navigation is also based on an extracted local model of the corridor, the system is quite robust to people traversing the corridor. People are another source of uncertainty in actions and observations, which is successfully treated by the probabilistic transition and observation models. Regarding doors state, the learning module adapts the probabilities to its real state, making the system more robust to this dynamic aspect of the environment. In order to improve the navigation capabilities of the proposed system, we are working on several future work lines. The first one is to enlarge the action and observation sets to navigate in more complex or generic environments. For example, to traverse large halls or unstructured areas, a "wall-following" or "trajectory-following" action would be useful. Besides, we are also working on the incorporation of new observations from new sensors, such as a compass (to discriminate the four orientations of the graph) and a wireless signal strength sensor. Enlarging the model doesn't affect the proposed global navigation algorithms. Regarding the learning system, future work is focused on automatically learning the POMDP structure from real data, making even easier the installation process. Another current research lines are the extension of localization, planning and learning probabilistic algorithms to multi-robot cooperative systems (SIMCA project) and the use of hierarchical topological models to expand the navigation system to larger structured environments In this chapter the authors have used a single non?standard mathematical framework, the Conformal Geometric Algebra, in order to simplify the set of data structures that we usually use with the traditional methods. The key idea is to define and use a set of products in CGA that will be enough to generate conformal transformations, manifolds as ruled surfaces and develop incidence algebra operations, as well as solve equations and obtain directed distances between different kinds of geometric primitives. Thus, within this approach, all those different mathematical entities and tasks can be done simultaneously, without the necessity of abandoning the system. Using conformal geometric algebra we even show that it is possible to find three grasping points for each kind of object, based on the intrinsic information of the object. The hand`s kinematic and the object structure can be easily related to each other in order to manage a natural and feasible grasping where force equilibrium is always guaranteed. These are only some applications that could show to the robotic and computer vision communities the useful insights and advantages of the CGA, and we invite them to adopt, explore and implement new tasks with this novel framework, expanding its horizon to new possibilities for robots equipped with stereo systems, range data, laser, omnidirectional and odometry