Improving on-line Monte Carlo POMDP (DESTOP in particular) in discrete spaces through the use of importance sampling, and a nice summary of the problem and of current on-line POMDP approaches

Luo, Y., Bai, H., Hsu, D., & Lee, W. S., Importance sampling for online planning under uncertainty, The International Journal of Robotics Research, 38(2–3), 162–181, 2019 DOI: 10.1177/0278364918780322.

The partially observable Markov decision process (POMDP) provides a principled general framework for robot planning under uncertainty. Leveraging the idea of Monte Carlo sampling, recent POMDP planning algorithms have scaled up to various challenging robotic tasks, including, real-time online planning for autonomous vehicles. To further improve online planning performance, this paper presents IS-DESPOT, which introduces importance sampling to DESPOT, a state-of-the-art sampling-based POMDP algorithm for planning under uncertainty. Importance sampling improves DESPOT’s performance when there are critical, but rare events, which are difficult to sample. We prove that IS-DESPOT retains the theoretical guarantee of DESPOT. We demonstrate empirically that importance sampling significantly improves the performance of online POMDP planning for suitable tasks. We also present a general method for learning the importance sampling distribution.

On the use of ubiquous supercomputing for robotics

Leonardo Camargo-Forero, Pablo Royo, Xavier Prats, The ARCHADE: Ubiquitous Supercomputing for robotics. Part I: Philosophy, Robotics and Autonomous Systems, Volume 114, 2019, Pages 187-198 DOI: 10.1016/j.robot.2019.01.006.

In this work, we introduce Ubiquitous Supercomputing for robotics with the objective of opening our imagination to the development of new powerful heterogeneous multi-robot systems able to perform all kind of missions. Supercomputing, also known as High Performance computing (HPC) is the tool that allows us to predict the weather, understand the origins of the universe, create incredibly realistic fantasy movies, send personalized advertisement to millions of users worldwide and much more. Robotics has been mostly absent in its use of HPC but some previous works have lightly flirted with it. With the findings presented in here, we propose a ubiquitous supercomputing ontology, which allows describing systems made up of robots, traditional HPC infrastructures, sensors, actuators and people and exhibiting scalability, user-transparency and ultimately higher computing efficiency. Moreover, we present a technology called The ARCHADE, which facilitates the development, implementation and operation of such systems, and we propose a mechanism to define and automatize missions carried out by ubiquitous supercomputing systems. As a proof of concept, we present a system depicted as Tigers VS Hunters, which illustrates the potential of this technology. The results presented in here are part of a two series work introducing The ARCHADE. This first delivery presents its philosophy and main features. Correspondingly the second part will present a set of use cases and a complete performance benchmark. Supercomputing is part of our lives and it can be found in many research and industrial endeavors. With the ubiquitous supercomputing ontology and The ARCHADE, supercomputing will become part of robotics as well, bringing it therefore everywhere.

RL and Inverse RL based on MDPs for autonomous vehicles, and a nice historical review of the topic of a.v.

Changxi You, Jianbo Lu, Dimitar Filev, Panagiotis Tsiotras, Advanced planning for autonomous vehicles using reinforcement learning and deep inverse reinforcement learning, Robotics and Autonomous Systems, Volume 114, 2019, Pages 1-18 DOI: 10.1016/j.robot.2019.01.003.

Autonomous vehicles promise to improve traffic safety while, at the same time, increase fuel efficiency and reduce congestion. They represent the main trend in future intelligent transportation systems. This paper concentrates on the planning problem of autonomous vehicles in traffic. We model the interaction between the autonomous vehicle and the environment as a stochastic Markov decision process (MDP) and consider the driving style of an expert driver as the target to be learned. The road geometry is taken into consideration in the MDP model in order to incorporate more diverse driving styles. The desired, expert-like driving behavior of the autonomous vehicle is obtained as follows: First, we design the reward function of the corresponding MDP and determine the optimal driving strategy for the autonomous vehicle using reinforcement learning techniques. Second, we collect a number of demonstrations from an expert driver and learn the optimal driving strategy based on data using inverse reinforcement learning. The unknown reward function of the expert driver is approximated using a deep neural-network (DNN). We clarify and validate the application of the maximum entropy principle (MEP) to learn the DNN reward function, and provide the necessary derivations for using the maximum entropy principle to learn a parameterized feature (reward) function. Simulated results demonstrate the desired driving behaviors of an autonomous vehicle using both the reinforcement learning and inverse reinforcement learning techniques.

Comparison of map-matching methods

Héber Sobreira, Carlos M. Costa, Ivo Sousa, Luis Rocha, José Lima, P. C. M. A. Farias, Paulo Costa, A. Paulo Moreira, Map-Matching Algorithms for Robot Self-Localization: A Comparison Between Perfect Match, Iterative Closest Point and Normal Distributions Transform, Journal of Intelligent & Robotic Systems, March 2019, Volume 93, Issue 3–4, pp 533–546 DOI: 10.1007/s10846-017-0765-5.

The self-localization of mobile robots in the environment is one of the most fundamental problems in the robotics navigation field. It is a complex and challenging problem due to the high requirements of autonomous mobile vehicles, particularly with regard to the algorithms accuracy, robustness and computational efficiency. In this paper, we present a comparison of three of the most used map-matching algorithms applied in localization based on natural landmarks: our implementation of the Perfect Match (PM) and the Point Cloud Library (PCL) implementation of the Iterative Closest Point (ICP) and the Normal Distribution Transform (NDT). For the purpose of this comparison we have considered a set of representative metrics, such as pose estimation accuracy, computational efficiency, convergence speed, maximum admissible initialization error and robustness to the presence of outliers in the robots sensors data. The test results were retrieved using our ROS natural landmark public dataset, containing several tests with simulated and real sensor data. The performance and robustness of the Perfect Match is highlighted throughout this article and is of paramount importance for real-time embedded systems with limited computing power that require accurate pose estimation and fast reaction times for high speed navigation. Moreover, we added to PCL a new algorithm for performing correspondence estimation using lookup tables that was inspired by the PM approach to solve this problem. This new method for computing the closest map point to a given sensor reading proved to be 40 to 60 times faster than the existing k-d tree approach in PCL and allowed the Iterative Closest Point algorithm to perform point cloud registration 5 to 9 times faster.

Models of brain based on artificial neural networks

James C.R. Whittington, Rafal Bogacz, Theories of Error Back-Propagation in the Brain, Trends in Cognitive Sciences, Volume 23, Issue 3, 2019, Pages 235-250 DOI: 10.1016/j.tics.2018.12.005.

This review article summarises recently proposed theories on how neural circuits in the brain could approximate the error back-propagation algorithm used by artificial neural networks. Computational models implementing these theories achieve learning as efficient as artificial neural networks, but they use simple synaptic plasticity rules based on activity of presynaptic and postsynaptic neurons. The models have similarities, such as including both feedforward and feedback connections, allowing information about error to propagate throughout the network. Furthermore, they incorporate experimental evidence on neural connectivity, responses, and plasticity. These models provide insights on how brain networks might be organised such that modification of synaptic weights on multiple levels of cortical hierarchy leads to improved performance on tasks.

Model-based RL for controling a soft manipulator arm

T. G. Thuruthel, E. Falotico, F. Renda and C. Laschi, Model-Based Reinforcement Learning for Closed-Loop Dynamic Control of Soft Robotic Manipulators, IEEE Transactions on Robotics, vol. 35, no. 1, pp. 124-134, Feb. 2019. DOI: 10.1109/TRO.2018.2878318.

Dynamic control of soft robotic manipulators is an open problem yet to be well explored and analyzed. Most of the current applications of soft robotic manipulators utilize static or quasi-dynamic controllers based on kinematic models or linearity in the joint space. However, such approaches are not truly exploiting the rich dynamics of a soft-bodied system. In this paper, we present a model-based policy learning algorithm for closed-loop predictive control of a soft robotic manipulator. The forward dynamic model is represented using a recurrent neural network. The closed-loop policy is derived using trajectory optimization and supervised learning. The approach is verified first on a simulated piecewise constant strain model of a cable driven under-actuated soft manipulator. Furthermore, we experimentally demonstrate on a soft pneumatically actuated manipulator how closed-loop control policies can be derived that can accommodate variable frequency control and unmodeled external loads.

Selecting the best visual cues in the next future for reducing the computational cost of localization under limited computational resources

L. Carlone and S. Karaman, Attention and Anticipation in Fast Visual-Inertial Navigation, IEEE Transactions on Robotics, vol. 35, no. 1, pp. 1-20, Feb. 2019 DOI: 10.1109/TRO.2018.2872402.

We study a visual-inertial navigation (VIN) problem in which a robot needs to estimate its state using an on-board camera and an inertial sensor, without any prior knowledge of the external environment. We consider the case in which the robot can allocate limited resources to VIN, due to tight computational constraints. Therefore, we answer the following question: under limited resources, what are the most relevant visual cues to maximize the performance of VIN? Our approach has four key ingredients. First, it is task-driven, in that the selection of the visual cues is guided by a metric quantifying the VIN performance. Second, it exploits the notion of anticipation, since it uses a simplified model for forward-simulation of robot dynamics, predicting the utility of a set of visual cues over a future time horizon. Third, it is efficient and easy to implement, since it leads to a greedy algorithm for the selection of the most relevant visual cues. Fourth, it provides formal performance guarantees: we leverage submodularity to prove that the greedy selection cannot be far from the optimal (combinatorial) selection. Simulations and real experiments on agile drones show that our approach ensures state-of-the-art VIN performance while maintaining a lean processing time. In the easy scenarios, our approach outperforms appearance-based feature selection in terms of localization errors. In the most challenging scenarios, it enables accurate VIN while appearance-based feature selection fails to track robot’s motion during aggressive maneuvers.

An application of MDPs to UAV collision-free navigation with an interesting taxonomy of the state-of-the-art

Xiang Yu1, Xiaobin Zhou2, Youmin Zhang, Collision-Free Trajectory Generation and Tracking for UAVs Using Markov Decision Process in a Cluttered Environment, Journal of Intelligent & Robotic Systems, 2019, 93:17–32 DOI: 10.1007/s10846-018-0802-z.

A collision-free trajectory generation and tracking method capable of re-planning unmanned aerial vehicle (UAV) trajectories can increase flight safety and decrease the possibility of mission failures. In this paper, a Markov decision process (MDP) based algorithm combined with backtracking method is presented to create a safe trajectory in the case of hostile environments. Subsequently, a differential flatness method is adopted to smooth the profile of the rerouted trajectory for satisfying the UAV physical constraints. Lastly, a flight controller based on passivity-based control (PBC) is designed to maintain UAV’s stability and trajectory tracking performance. simulation results demonstrate that the UAV with the proposed strategy is capable of avoiding obstacles in a hostile environment.

Interesting mathematical study of the properties of graphs for graph-based SLAM and other graph-based estimation problems

Khosoussi, K., Giamou, M., Sukhatme, G. S., Huang, S., Dissanayake, G., & How, J. P., Reliable Graphs for SLAM, The International Journal of Robotics Research, 2019, DOI: 10.1177/0278364918823086.

Estimation-over-graphs (EoG) is a class of estimation problems that admit a natural graphical representation. Several key problems in robotics and sensor networks, including sensor network localization, synchronization over a group, and simultaneous localization and mapping (SLAM) fall into this category. We pursue two main goals in this work. First, we aim to characterize the impact of the graphical structure of SLAM and related problems on estimation reliability. We draw connections between several notions of graph connectivity and various properties of the underlying estimation problem. In particular, we establish results on the impact of the weighted number of spanning trees on the D-optimality criterion in 2D SLAM. These results enable agents to evaluate estimation reliability based only on the graphical representation of the EoG problem. We then use our findings and study the problem of designing sparse SLAM problems that lead to reliable maximum likelihood estimates through the synthesis of sparse graphs with the maximum weighted tree connectivity. Characterizing graphs with the maximum number of spanning trees is an open problem in general. To tackle this problem, we establish several new theoretical results, including the monotone log-submodularity of the weighted number of spanning trees. We exploit these structures and design a complementary greedy–convex pair of efficient approximation algorithms with provable guarantees. The proposed synthesis framework is applied to various forms of the measurement selection problem in resource-constrained SLAM. Our algorithms and theoretical findings are validated using random graphs, existing and new synthetic SLAM benchmarks, and publicly available real pose-graph SLAM datasets.

SLAM based on submap joining that achieves linear cost through a novel choice of the reference frame of each submap, and an interesting related works on map joining, i.e., considering submaps as observations

Liang Zhao, Shoudong Huang, Gamini Dissanayake, Linear SLAM: Linearising the SLAM problems using submap joining, Automatica, Volume 100, 2019, Pages 231-246, DOI: 10.1016/j.automatica.2018.10.037.

The main contribution of this paper is a new submap joining based approach for solving large-scale Simultaneous Localization and Mapping (SLAM) problems. Each local submap is independently built using the local information through solving a small-scale SLAM; the joining of submaps mainly involves solving linear least squares and performing nonlinear coordinate transformations. Through approximating the local submap information as the state estimate and its corresponding information matrix, judiciously selecting the submap coordinate frames, and approximating the joining of a large number of submaps by joining only two maps at a time, either sequentially or in a more efficient Divide and Conquer manner, the nonlinear optimization process involved in most of the existing submap joining approaches is avoided. Thus the proposed submap joining algorithm does not require initial guess or iterations since linear least squares problems have closed-form solutions. The proposed Linear SLAM technique is applicable to feature-based SLAM, pose graph SLAM and D-SLAM, in both two and three dimensions, and does not require any assumption on the character of the covariance matrices. Simulations and experiments are performed to evaluate the proposed Linear SLAM algorithm. Results using publicly available datasets in 2D and 3D show that Linear SLAM produces results that are very close to the best solutions that can be obtained using full nonlinear optimization algorithm started from an accurate initial guess. The C/C++ and MATLAB source codes of Linear SLAM are available on OpenSLAM.