In this paper, we propose a new strategy for learning inertial robotic navigation models. The proposed strategy enhances the generalisability of end-to-end inertial modelling, and is aimed at wheeled robotic deployments. Concretely, the paper describes the following. (1) Using precision robotics, we empirically characterise the effect of changing the sensor position during navigation on the distribution of raw inertial signals, as well as the corresponding impact on learnt latent spaces. (2) We propose neural architectures and algorithms to assimilate knowledge from an indexed set of sensor positions in order to enhance the robustness and generalisability of robotic inertial tracking in the field. Our scheme of choice uses continuous domain adaptation (DA) and optimal transport (OT). (3) In our evaluation, continuous OT DA outperforms a continuous adversarial DA baseline, while also showing quantifiable learning benefits over simple data augmentation. We will release our dataset to help foster future research.
Breakthroughs in machine learning in the last decade have led to `digital intelligence', i.e. machine learning models capable of learning from vast amounts of labeled data to perform several digital tasks such as speech recognition, face recognition, machine translation and so on. The goal of this thesis is to make progress towards designing algorithms capable of `physical intelligence', i.e. building intelligent autonomous navigation agents capable of learning to perform complex navigation tasks in the physical world involving visual perception, natural language understanding, reasoning, planning, and sequential decision making. Despite several advances in classical navigation methods in the last few decades, current navigation agents struggle at long-term semantic navigation tasks. In the first part of the thesis, we discuss our work on short-term navigation using end-to-end reinforcement learning to tackle challenges such as obstacle avoidance, semantic perception, language grounding, and reasoning. In the second part, we present a new class of navigation methods based on modular learning and structured explicit map representations, which leverage the strengths of both classical and end-to-end learning methods, to tackle long-term navigation tasks. We show that these methods are able to effectively tackle challenges such as localization, mapping, long-term planning, exploration and learning semantic priors. These modular learning methods are capable of long-term spatial and semantic understanding and achieve state-of-the-art results on various navigation tasks.
With the rise of automation, unmanned vehicles became a hot topic both as commercial products and as a scientific research topic. It composes a multi-disciplinary field of robotics that encompasses embedded systems, control theory, path planning, Simultaneous Localization and Mapping (SLAM), scene reconstruction, and pattern recognition. In this work, we present our exploratory research of how sensor data fusion and state-of-the-art machine learning algorithms can perform the Embodied Artificial Intelligence (E-AI) task called Visual Semantic Navigation. This task, a.k.a Object-Goal Navigation (ObjectNav) consists of autonomous navigation using egocentric visual observations to reach an object belonging to the target semantic class without prior knowledge of the environment. Our method reached fourth place on the Habitat Challenge 2021 ObjectNav on the Minival phase and the Test-Standard Phase.
Mobile robots are desired with resilience to safely interact with prior-unknown environments and finally accomplish given tasks. This paper utilizes instantaneous local sensory data to stimulate the safe feedback motion planning (SFMP) strategy with adaptability to diverse prior-unknown environments without building a global map. This is achieved by the numerical optimization with the constraints, referred to as instantaneous local control barrier functions (IL-CBFs) and goal-driven control Lyapunov functions (GD-CLFs), learned from perceptional signals. In particular, the IL-CBFs reflecting potential collisions and GD-CLFs encoding incrementally discovered subgoals are first online learned from local perceptual data. Then, the learned IL-CBFs are united with GD-CLFs in the context of quadratic programming (QP) to generate the safe feedback motion planning strategy. Rather importantly, an optimization over the admissible control space of IL-CBFs is conducted to enhance the solution feasibility of QP. The SFMP strategy is developed with theoretically guaranteed collision avoidance and convergence to destinations. Numerical simulations are conducted to reveal the effectiveness of the proposed SFMP strategy that drives mobile robots to safely reach the destination incrementally in diverse prior-unknown environments.
Pedestrian motion behavior involves a combination of individual goals and social interactions with other agents. In this article, we present an asymmetrical bidirectional recurrent neural network architecture called U-RNN to encode pedestrian trajectories and evaluate its relevance to replace LSTMs for various forecasting models. Experimental results on the Trajnet++ benchmark show that the U-LSTM variant yields better results regarding every available metrics (ADE, FDE, Collision rate) than common trajectory encoders for a variety of approaches and interaction modules, suggesting that the proposed approach is a viable alternative to the de facto sequence encoding RNNs. Our implementation of the asymmetrical Bi-RNNs for the Trajnet++ benchmark is available at: github.com/JosephGesnouin/Asymmetrical-Bi-RNNs-to-encode-pedestrian-trajectories
Estimating position and orientation change of a mobile platform from two consecutive point clouds provided by a high-resolution sensor is a key problem in autonomous navigation. In particular, scan matching algorithms aim to find the translation and rotation of the platform such that the two point clouds coincide. The association of measurements in point cloud one with measurements in point cloud two is a problem inherent to scan matching. Existing methods perform non-probabilistic data association, i.e., they assume a single association hypothesis. This leads to overconfident pose estimates and reduced estimation accuracy in ambiguous environments. Our probabilistic scan matching approach addresses this issue by considering all association hypotheses with their respective likelihoods. We formulate a holistic Bayesian estimation problem for both data association and pose estimation and present the corresponding joint factor graph. Near-optimum maximum a posteriori (MAP) estimates of the sensor pose are computed by performing iterative message passing on the factor graph. Our numerical study shows performance improvements compared to non-probabilistic scan matching methods that are based on the normal distributions transform (NDT) and implicit moving least squares (IMLS).
Recent advances in deep learning have led to the development of accurate and efficient models for various computer vision applications such as classification, segmentation, and detection. However, learning highly accurate models relies on the availability of large-scale annotated datasets. Due to this, model performance drops drastically when evaluated on label-scarce datasets having visually distinct images, termed as domain adaptation problem. There is a plethora of works to adapt classification and segmentation models to label-scarce target datasets through unsupervised domain adaptation. Considering that detection is a fundamental task in computer vision, many recent works have focused on developing novel domain adaptive detection techniques. Here, we describe in detail the domain adaptation problem for detection and present an extensive survey of the various methods. Furthermore, we highlight strategies proposed and the associated shortcomings. Subsequently, we identify multiple aspects of the problem that are most promising for future research. We believe that this survey shall be valuable to the pattern recognition experts working in the fields of computer vision, biometrics, medical imaging, and autonomous navigation by introducing them to the problem, and familiarizing them with the current status of the progress while providing promising directions for future research.
Accurate reconstruction of static environments from LiDAR scans of scenes containing dynamic objects, which we refer to as Dynamic to Static Translation (DST), is an important area of research in Autonomous Navigation. This problem has been recently explored for visual SLAM, but to the best of our knowledge no work has been attempted to address DST for LiDAR scans. The problem is of critical importance due to wide-spread adoption of LiDAR in Autonomous Vehicles. We show that state-of the art methods developed for the visual domain when adapted for LiDAR scans perform poorly. We develop DSLR, a deep generative model which learns a mapping between dynamic scan to its static counterpart through an adversarially trained autoencoder. Our model yields the first solution for DST on LiDAR that generates static scans without using explicit segmentation labels. DSLR cannot always be applied to real world data due to lack of paired dynamic-static scans. Using Unsupervised Domain Adaptation, we propose DSLR-UDA for transfer to real world data and experimentally show that this performs well in real world settings. Additionally, if segmentation information is available, we extend DSLR to DSLR-Seg to further improve the reconstruction quality. DSLR gives the state of the art performance on simulated and real-world datasets and also shows at least 4x improvement. We show that DSLR, unlike the existing baselines, is a practically viable model with its reconstruction quality within the tolerable limits for tasks pertaining to autonomous navigation like SLAM in dynamic environments.
In this work we showcase the design and assessment of the performance of a multi-camera UAV, when coupled with state-of-the-art planning and mapping algorithms for autonomous navigation. The system leverages state-of-the-art receding horizon exploration techniques for Next-Best-View (NBV) planning with 3D and semantic information, provided by a reconfigurable multi stereo camera system. We employ our approaches in an autonomous drone-based inspection task and evaluate them in an autonomous exploration and mapping scenario. We discuss the advantages and limitations of using multi stereo camera flying systems, and the trade-off between number of cameras and mapping performance.
While imitation learning for vision based autonomous mobile robot navigation has recently received a great deal of attention in the research community, existing approaches typically require state action demonstrations that were gathered using the deployment platform. However, what if one cannot easily outfit their platform to record these demonstration signals or worse yet the demonstrator does not have access to the platform at all? Is imitation learning for vision based autonomous navigation even possible in such scenarios? In this work, we hypothesize that the answer is yes and that recent ideas from the Imitation from Observation (IfO) literature can be brought to bear such that a robot can learn to navigate using only ego centric video collected by a demonstrator, even in the presence of viewpoint mismatch. To this end, we introduce a new algorithm, Visual Observation only Imitation Learning for Autonomous navigation (VOILA), that can successfully learn navigation policies from a single video demonstration collected from a physically different agent. We evaluate VOILA in the photorealistic AirSim simulator and show that VOILA not only successfully imitates the expert, but that it also learns navigation policies that can generalize to novel environments. Further, we demonstrate the effectiveness of VOILA in a real world setting by showing that it allows a wheeled Jackal robot to successfully imitate a human walking in an environment using a video recorded using a mobile phone camera.
The ability to simultaneously track and reconstruct multiple objects moving in the scene is of the utmost importance for robotic tasks such as autonomous navigation and interaction. Virtually all of the previous attempts to map multiple dynamic objects have evolved to store individual objects in separate reconstruction volumes and track the relative pose between them. While simple and intuitive, such formulation does not scale well with respect to the number of objects in the scene and introduces the need for an explicit occlusion handling strategy. In contrast, we propose a map representation that allows maintaining a single volume for the entire scene and all the objects therein. To this end, we introduce a novel multi-object TSDF formulation that can encode multiple object surfaces at any given location in the map. In a multiple dynamic object tracking and reconstruction scenario, our representation allows maintaining accurate reconstruction of surfaces even while they become temporarily occluded by other objects moving in their proximity. We evaluate the proposed TSDF++ formulation on a public synthetic dataset and demonstrate its ability to preserve reconstructions of occluded surfaces when compared to the standard TSDF map representation.
In this paper, we develop a safe decision-making method for self-driving cars in a multi-lane, single-agent setting. The proposed approach utilizes deep reinforcement learning (RL) to achieve a high-level policy for safe tactical decision-making. We address two major challenges that arise solely in autonomous navigation. First, the proposed algorithm ensures that collisions never happen, and therefore accelerate the learning process. Second, the proposed algorithm takes into account the unobservable states in the environment. These states appear mainly due to the unpredictable behavior of other agents, such as cars, and pedestrians, and make the Markov Decision Process (MDP) problematic when dealing with autonomous navigation. Simulations from a well-known self-driving car simulator demonstrate the applicability of the proposed method
LiDAR odometry plays an important role in self-localization and mapping for autonomous navigation, which is usually treated as a scan registration problem. Although having achieved promising performance on KITTI odometry benchmark, the conventional searching tree-based approach still has the difficulty in dealing with the large scale point cloud efficiently. The recent spherical range image-based method enjoys the merits of fast nearest neighbor search by spherical mapping. However, it is not very effective to deal with the ground points nearly parallel to LiDAR beams. To address these issues, we propose a novel efficient LiDAR odometry approach by taking advantage of both non-ground spherical range image and bird's-eye-view map for ground points. Moreover, a range adaptive method is introduced to robustly estimate the local surface normal. Additionally, a very fast and memory-efficient model update scheme is proposed to fuse the points and their corresponding normals at different time-stamps. We have conducted extensive experiments on KITTI odometry benchmark, whose promising results demonstrate that our proposed approach is effective.
We describe a robotic learning system for autonomous exploration and navigation in diverse, open-world environments. At the core of our method is a learned latent variable model of distances and actions, along with a non-parametric topological memory of images. We use an information bottleneck to regularize the learned policy, giving us (i) a compact visual representation of goals, (ii) improved generalization capabilities, and (iii) a mechanism for sampling feasible goals for exploration. Trained on a large offline dataset of prior experience, the model acquires a representation of visual goals that is robust to task-irrelevant distractors. We demonstrate our method on a mobile ground robot in open-world exploration scenarios. Given an image of a goal that is up to 80 meters away, our method leverages its representation to explore and discover the goal in under 20 minutes, even amidst previously-unseen obstacles and weather conditions. Please check out the project website for videos of our experiments and information about the real-world dataset used at https://sites.google.com/view/recon-robot.
Provision of some autonomous functions to an agricultural vehicle would lighten the job of the operator but in doing so, the accuracy should not be lost to still obtain an optimal yield. Autonomous navigation of an agricultural vehicle involves the control of different dynamic subsystems, such as the yaw angle dynamics and the longitudinal speed dynamics. In this study, a proportional-integral-derivative controller is used to control the longitudinal velocity of the tractor. For the control of the yaw angle dynamics, a proportional-derivative controller works in parallel with a type-2 fuzzy neural network. In such an arrangement, the former ensures the stability of the related subsystem, while the latter learns the system dynamics and becomes the leading controller. In this way, instead of modeling the interactions between the subsystems prior to the design of model-based control, we develop a control algorithm which learns the interactions online from the measured feedback error. In addition to the control of the stated subsystems, a kinematic controller is needed to correct the errors in both the x- and the y- axis for the trajectory tracking problem of the tractor. To demonstrate the real-time abilities of the proposed control scheme, an autonomous tractor is equipped with the use of reasonably priced sensors and actuators. Experimental results show the efficacy and efficiency of the proposed learning algorithm.
Recently, mobile robots have become important tools in various industries, especially in logistics. Deep reinforcement learning emerged as an alternative planning method to replace overly conservative approaches and promises more efficient and flexible navigation. However, deep reinforcement learning approaches are not suitable for long-range navigation due to their proneness to local minima and lack of long term memory, which hinders its widespread integration into industrial applications of mobile robotics. In this paper, we propose a navigation system incorporating deep-reinforcement-learning-based local planners into conventional navigation stacks for long-range navigation. Therefore, a framework for training and testing the deep reinforcement learning algorithms along with classic approaches is presented. We evaluated our deep-reinforcement-learning-enhanced navigation system against various conventional planners and found that our system outperforms them in terms of safety, efficiency and robustness.
The autonomous operation of flexible-wing aircraft is technically challenging and has never been presented within literature. The lack of an exact modeling framework is due to the complex nonlinear aerodynamic relationships governed by the deformations in the flexible-wing shape, which in turn complicates the controls and instrumentation setup of the navigation system. This urged for innovative approaches to interface affordable instrumentation platforms to autonomously control this type of aircraft. This work leverages ideas from instrumentation and measurements, machine learning, and optimization fields in order to develop an autonomous navigation system for a flexible-wing aircraft. A novel machine learning process based on a guiding search mechanism is developed to interface real-time measurements of wing-orientation dynamics into control decisions. This process is realized using an online value iteration algorithm that decides on two improved and interacting model-free control strategies in real-time. The first strategy is concerned with achieving the tracking objectives while the second supports the stability of the system. A neural network platform that employs adaptive critics is utilized to approximate the control strategies while approximating the assessments of their values. An experimental actuation system is utilized to test the validity of the proposed platform. The experimental results are shown to be aligned with the stability features of the proposed model-free adaptive learning approach.
Fitting concentric geometric objects to digitized data is an important problem in many areas such as iris detection, autonomous navigation, and industrial robotics operations. There are two common approaches to fitting geometric shapes to data: the geometric (iterative) approach and algebraic (non-iterative) approach. The geometric approach is a nonlinear iterative method that minimizes the sum of the squares of Euclidean distances of the observed points to the ellipses and regarded as the most accurate method, but it needs a good initial guess to improve the convergence rate. The algebraic approach is based on minimizing the algebraic distances with some constraints imposed on parametric space. Each algebraic method depends on the imposed constraint, and it can be solved with the aid of the generalized eigenvalue problem. Only a few methods in literature were developed to solve the problem of concentric ellipses. Here we study the statistical properties of existing methods by firstly establishing a general mathematical and statistical framework for this problem. Using rigorous perturbation analysis, we derive the variances and biasedness of each method under the small-sigma model. We also develop new estimators, which can be used as reliable initial guesses for other iterative methods. Then we compare the performance of each method according to their theoretical accuracy. Not only do our methods described here outperform other existing non-iterative methods, they are also quite robust against large noise. These methods and their practical performances are assessed by a series of numerical experiments on both synthetic and real data.
Obstacle avoidance is a fundamental and challenging problem for autonomous navigation of mobile robots. In this paper, we consider the problem of obstacle avoidance in simple 3D environments where the robot has to solely rely on a single monocular camera. In particular, we are interested in solving this problem without relying on localization, mapping, or planning techniques. Most of the existing work consider obstacle avoidance as two separate problems, namely obstacle detection, and control. Inspired by the recent advantages of deep reinforcement learning in Atari games and understanding highly complex situations in Go, we tackle the obstacle avoidance problem as a data-driven end-to-end deep learning approach. Our approach takes raw images as input and generates control commands as output. We show that discrete action spaces are outperforming continuous control commands in terms of expected average reward in maze-like environments. Furthermore, we show how to accelerate the learning and increase the robustness of the policy by incorporating predicted depth maps by a generative adversarial network.
For robots navigating using only a camera, illumination changes in indoor environments can cause re-localization failures during autonomous navigation. In this paper, we present a multi-session visual SLAM approach to create a map made of multiple variations of the same locations in different illumination conditions. The multi-session map can then be used at any hour of the day for improved re-localization capability. The approach presented is independent of the visual features used, and this is demonstrated by comparing re-localization performance between multi-session maps created using the RTAB-Map library with SURF, SIFT, BRIEF, BRISK, KAZE, DAISY and SuperPoint visual features. The approach is tested on six mapping and six localization sessions recorded at 30 minute intervals during sunset using a Google Tango phone in a real apartment.