Abstract-Both landmark measurements and loop closures are used to correct for odometry drift in SLAM solutions. However, if any of the measurements are incorrect (e.g. due to perceptual aliasing) standard SLAM algorithms fail catastrophically and can not return an accurate map. A number of algorithms have been proposed that are robust to loop closure errors, but it is shown in this paper that they can not provide robust solutions when landmark measurement errors occur. The root cause of the problem is that most robust SLAM algorithms only focus on creating a locally consistent map (by evaluating whether measurements appear correct individually) rather than a globally consistent map. This paper proposes a new formulation of the robust SLAM problem that explicitly requires finding a globally consistent solution. Motivated by the new cost function, a novel incremental SLAM algorithm is developed that provides accurate solutions for datasets with landmark or loop closure measurement errors. Simulated and experimental results of the new algorithm, called incremental SLAM with consistency-checking, show that the new algorithm provides significantly more accurate results than state-of-theart robust SLAM methods for datasets with incorrect landmark measurements and can match the performance of current robust SLAM methods for datasets with incorrect loop closures.

Abstract-Long-term operations of resource-constrained robots typically require hard decisions be made about which data to process and/or retain. The question then arises of how to choose which data is most useful to keep to achieve the task at hand. As spacial scale grows, the size of the map will grow without bound, and as temporal scale grows, the number of measurements will grow without bound. In this work, we present the first known approach to tackle both of these issues.The approach has two stages. First, a subset of the variables (focused variables) is selected that are most useful for a particular task. Second, a task-agnostic and principled method (focused inference) is proposed to select a subset of the measurements that maximizes the information over the focused variables. The approach is then applied to the specific task of robot navigation in an obstacle-laden environment. A landmark selection method is proposed to minimize the probability of collision and then select the set of measurements that best localizes those landmarks. It is shown that the two-stage approach outperforms both only selecting measurement and only selecting landmarks in terms of minimizing the probability of collision. The performance improvement is validated through detailed simulation and real experiments on a Pioneer robot.

One of the major challenges for state estimation algorithms, such as the Kalman lter, is the impact of outliers that do not match the assumed Gaussian process and measurement noise. When these errors occur they can induce large state estimate errors and even lter divergence. This paper presents a robust recursive ltering algorithm, the l 1 -norm lter, that can provide reliable state estimates in the presence of both measurement and state propagation outliers. The algorithm consists of a convex optimization to detect the outliers followed by a state update step based on the results of the error detection. Monte Carlo simulation results are presented to demonstrate the robustness of the l 1 -norm lter estimates to both state prediction and measurement outliers. Finally, vision-aided navigation experimental results are presented that demonstrate that the proposed algorithm can provide improved state estimation performance over existing robust ltering approaches.

No abstract