Abstract

Most methods that attempt to tackle the problem of Autonomous Driving and overtaking usually try to either directly minimize an objective function or iteratively in a Reinforcement Learning like framework to generate motor actions given a set of inputs. We follow a similar trend but train the agent in a way similar to a curriculum learning approach where the agent is first given an easier problem to solve, followed by a harder problem. We use Deep Deterministic Policy Gradients to learn overtaking maneuvers for a car, in presence of multiple other cars, in a simulated highway scenario. The novelty of our approach lies in the training strategy used where we teach the agent to drive in a manner similar to the way humans learn to drive and the fact that our reward function uses only the raw sensor data at the current time step. This method, which resembles a curriculum learning approach is able to learn smooth maneuvers, largely collision free, wherein the agent overtakes all other cars, independent of the track and number of cars in the scene.

Abstract

Accurate localization of other traffic participants is a vital task in autonomous driving systems. State-of-the-art systems employ a combination of sensing modalities such as RGB cameras and LiDARs for localizing traffic participants, but most such emonstrations have been confined to plain roads. We demonstrate, to the best of our knowledge, the first results for monocular object localization and shape estimation on surfaces that do not share the same plane with the moving monocular camera. We approximate road surfaces by local planar patches and use semantic cues from vehicles in the scene to initialize a local bundle-adjustment like procedure that simultaneously estimates the pose and shape of the vehicles, and the orientation of the local ground plane on which the vehicle stands as well. We evaluate the proposed approach on the KITTI and SYNTHIA-SF benchmarks, for a variety of road plane configurations. The proposed approach significantly improves the state-of-the-art for monocular object localization on arbitrarily-shaped roads.

Abstract

Objects in space often exhibit a tumbling motion around the major inertial axis. In this paper, we address the image based visual servoing of a robotic system towards an uncooperative tumbling object. In contrast to previous approaches that require explicit reconstruction of the object and an estimation of its velocity, we propose a novel controller that is able to minimize the feature error directly in image space. This is achieved by observing that the feature points on the tumbling object follow a circular path around the axis of rotation and their projection creates an elliptical track in the image plane. Our controller minimizes the error between this elliptical track and the desired features, such that at the desired pose the features lie on the circumference of the ellipse. The effectiveness of our framework is exhibited by implementing the algorithm in simulation as well on a mobile robot.

Abstract

3D LiDARs and 2D cameras are increasingly being used alongside each other in sensor rigs for perception tasks. Before these sensors can be used to gather meaningful data, however, their extrinsics (and intrinsics) need to be accurately calibrated, as the performance of the sensor rig is extremely sensitive to these calibration parameters. A vast majority of existing calibration techniques require significant amounts of data and/or calibration targets and human effort, severely impacting their applicability in large-scale production systems. We address this gap with CalibNet: a geometrically supervised deep network capable of automatically estimating the 6-DoF rigid body transformation between a 3D LiDAR and a 2D camera in real-time. CalibNet alleviates the need for calibration targets, thereby resulting in significant savings in calibration efforts. During training, the network only takes as input a LiDAR point cloud, the corresponding monocular image, and the camera calibration matrix K. At train time, we do not impose direct supervision (i.e., we do not directly regress to the calibration parameters, for example). Instead, we train the network to predict calibration parameters that maximize the geometric and photometric consistency of the input images and point clouds. CalibNet learns to iteratively solve the underlying geometric problem and accurately predicts extrinsic calibration parameters for a wide range of mis-calibrations, without requiring retraining or domain adaptation.

Abstract

Sparse Bayesian learning (SBL) has been used to obtain source direction-of-arrivals (DoAs) from uniform linear array (ULA) data. The maximum number of sources that can be resolved using a ULA is limited by the number of sensors in the array. It is known that sparse linear arrays such as co-prime and nested arrays can resolve more sources than the number of sensors. In this paper we demonstrate this using SBL. We compute the mean squared error in source power estimation as various parameters are varied.

Abstract

We propose a novel Line based parameterization for category specific CAD models. The proposed parameterization associates 3D category-specific CAD model and object under consideration using a dictionary based RANSAC method that uses object Viewpoints as prior and edges detected in the respective intensity image of the scene. The association problem is posed as a classical Geometry problem rather than being dataset driven, thus saving the time and labour that one invests in annotating dataset to train Keypoint Network[1, 2] for different category objects. Besides eliminating the need of dataset preparation, the approach also speeds up the entire process as this method processes the image only once for all objects, thus eliminating the need of invoking the network for every object in an image across all images. A 3D-2D edge association module followed by a resection algorithm for lines is used to recover object poses. The formulation optimizes for shape and pose of the object, thus aiding in recovering object 3D structure more accurately. Finally, a Factor Graph formulation is used to combine object poses with camera odometry to formulate a SLAM problem.

Abstract

Using simulations and data, we localize a quiet source in the presence of an interferer. The SWellEx-96 Event S59 consists of a submerged source towed along an isobath over a 65 min duration with an interferer traversing the source track. This range independent, multi-frequency scenario includes mismatch, non-stationary noise, and operational uncertainty. Mismatch is defined as a misalignment between the actual source field observed at the array and the modeled replica vector. The noise process changes likely with time. This is modelled as a heteroscedastic Gaussian process, meaning that the noise variance is non-stationary across snapshots. Sparse Bayesian learning (SBL) has been applied previously to the matched field processing application [Gemba et al, J. Acoust. Soc. Am., 141:3411-3420, 2017]. Results demonstrate that SBL exhibits desirable robustness properties and improved localization …

Abstract

Given a topology of local parity-check constraints, a maximally recoverable code (MRC) can correct all erasure patterns that are information-theoretically correctable. In a grid-like topology, there are a local constraints in every column forming a column code, b local constraints in every row forming a row code, and h global constraints in an (m × n) grid of codeword. Recently, Gopalan et al. initiated the study of MRCs under grid-like topology, and derived a necessary and sufficient condition, termed as the regularity condition, for an erasure pattern to be recoverable when a = 1, h = 0. In this paper, we consider MRCs for product topology ( h = 0). First, we construct a certain bipartite graph based on the erasure pattern satisfying the regularity condition for product topology (any a, b, h = 0) and show that there exists a complete matching in this graph. We then present an alternate direct proof of the sufficient condition when a = 1, h = 0. We later extend our technique to study the topology for a = 2, h = 0, and characterize a subset of recoverable erasure patterns in that case. For both a = 1, 2, our method of proof is uniform, i.e., by constructing tensor product G col ⊗ G row of generator matrices of column and row codes such that certain square sub-matrices retain full rank. The full-rank condition is proved by resorting to the matching identified earlier and also another set of matchings in erasure sub-patterns.

Abstract

Index coding for broadcast channels allows each receiver or client to retrieve its demanded message from its side information and the transmitted codeword. In general, a client may have to observe the entire codeword to decode its demanded message. However, downloading or querying the codeword symbols might involve costs at a client - such as network utilization costs and storage. Traditional index coding does not consider this client perspective, and as a result, for these codes the number of codeword symbols queried by a client per decoded message symbol, which we refer to as locality, could be large. In this paper we study a `client aware' approach to index coding by viewing the problem as a trade-off between the achievable broadcast rate and locality, where the objective is to minimize the rate for a given value of locality and vice versa. We first consider the minimum possible locality 1 and show that the …

Abstract

Most reinforcement learning algorithms are inefficient for learning multiple tasks in complex robotic systems, where different tasks share a set of actions. In such environments a compound policy may be learnt with shared neural network parameters, which performs multiple tasks concurrently. However such compound policy may get biased towards a task or the gradients from different tasks negate each other, making the learning unstable and sometimes less data efficient. In this paper, we propose a new approach for simultaneous training of multiple tasks sharing a set of common actions in continuous action spaces, which we call as DiGrad (Differential Policy Gradient). The proposed framework is based on differential policy gradients and can accommodate multi-task learning in a single actor-critic network. We also propose a simple heuristic in the differential policy gradient update to further improve the learning. The proposed architecture was tested on 8 link planar manipulator and 27 degrees of freedom (DoF) Humanoid for learning multi-goal reachability tasks for 3 and 2 end effectors respectively. We show that our approach supports efficient multi-task learning in complex robotic systems, outperforming related methods in continuous action spaces.