Abstract

Monocular SLAM is a well-studied problem and has shown significant progress in recent years, but still, challenges remain in creating a rich semantic description of the scene. Feature-based visual SLAMs are vulnerable to erroneous pose estimates due to insufficient tracking of mapped points or motion induced errors such as in large or in-place rotations. We present a new SLAM framework in which we use monocular edge based SLAM [1], along with category level models, to localize objects in the scene as well as improve the camera trajectory. In monocular SLAM systems, the camera track tends to break in conditions with abrupt motion which leads to reduction in the number of 2D point correspondences. In order to tackle this problem, we propose the first most principled formulation of its kind which integrates object category models in the core SLAM back-end to jointly optimize for the camera trajectory, object poses along with its shape and 3D structure. We show that our joint optimization is able to recover a better camera trajectory in such cases, as compared to Edge SLAM. Moreover, this method gives a better visualization incorporating object representations in the scene along with the 3D structure of the base SLAM system, which makes it useful for augmented reality (AR) applications

Abstract

In this paper, we present a novel control law for longitudinal speed control of autonomous vehicles. The key contributions of the proposed work include the design of a control law that reactively integrates the longitudinal surface gradient of the road into its operation. In contrast to the existing works, we found that integrating the path gradient into the control framework improves the speed tracking efficacy. Since the control law is implemented over a shrinking domain scheme, it minimizes the integrated error by recomputing the control inputs at every discretized step and consequently provides less reaction time. This makes our control law suitable for motion planning frameworks that are operating at high frequencies. Furthermore, our work is implemented using a generalized vehicle model and can be easily extended to other classes of vehicles. The performance of gradient aware - shrinking domain based controller is implemented and tested on an electric car. Results from the tests show the robustness of our control law for speed tracking efficiency on terrain with varying gradient while also considering stringent time constraints imposed by the planning framework.

Abstract

Multi-agent learning provides a potential framework for learning and simulating traffic behaviors. This paper proposes a novel architecture to learn multiple driving behaviors in a traffic scenario. The proposed architecture can learn multiple behaviors independently as well as simultaneously. We take advantage of the homogeneity of agents and learn in a parameter sharing paradigm. To further speed up the training process asynchronous updates are employed into the architecture. While learning different behaviors simultaneously, the given framework was also able to learn cooperation between the agents, without any explicit communication. We applied this framework to learn two important behaviors in driving: 1) Lane-Keeping and 2) Over-Taking. Results indicate faster convergence and learning of a more generic behavior, that is scalable to any number of agents. When compared the results with existing approaches, our results indicate equal and even better performance in some cases.

Abstract

Coded Caching is a promising solution to reduce the peak traffic in broadcast networks by prefetching the popular content close to end users and using coded transmissions. One of the chief issues of most coded caching schemes in literature is the issue of large subpacketization, i.e., they require each file to be divided into a large number of subfiles. In this work, we present a coded caching scheme using line graphs of bipartite graphs in conjunction with projective geometries over finite fields. The presented scheme achieves a rate Θ(K/log q K) (K being the number of users, q is some prime power) with subexponential subpacketization q O((logq K)^2) when cached fraction is upper bounded by a constant (M/N ≤ 1/q α , for some positive integer α). Compared to earlier schemes, the presented scheme has a lower subpacketization (albeit possessing a higher rate). We also present a new subpacketization …

Abstract

Multi-agent learning provides a potential framework for learning and simulating traffic behaviors. This paper proposes a novel architecture to learn multiple driving behaviors in a traffic scenario. The proposed architecture can learn multiple behaviors independently as well as simultaneously. We take advantage of the homogeneity of agents and learn in a parameter sharing paradigm. To further speed up the training process asynchronous updates are employed into the architecture. While learning different behaviors simultaneously, the given framework was also able to learn cooperation between the agents, without any explicit communication. We applied this framework to learn two important behaviors in driving: 1) Lane-Keeping and 2) Over-Taking. Results indicate faster convergence and learning of a more generic behavior, that is scalable to any number of agents. When compared the results with existing approaches, our results indicate equal and even better performance in some cases.

Abstract

Robots with joint or link compliance are gaining importance due to their inherent safety. In this paper four different designs of compliant shanks of an 8 degrees of freedom (DoF) biped robot are compared to get the least energy consumed during walk. Different leg and pelvis trajectories of the biped during walk are generated by varying the gait parameters. The deflection of the compliant links during walk is modeled using a finite element method (FEM). Simulations are carried out in which the optimal walking trajectory that consumes least energy is found by using a genetic algorithm (GA). Balance is ensured during walk by ensuring that the trajectory always satisfies the zero moment point (ZMP) criteria. The optimal trajectories obtained from the simulations were then experimentally evaluated for a compliant link biped robot. The motion of the different joints of the biped during the experiment was tracked using a vision system and the performance of the rigid link and compliant link bipeds compared. The best compliant link design was finally chosen from the four different designs.

Abstract

Coded Caching is a recent technique that optimizes the use of a multi-client broadcast channel by the use of local storage available at the clients and by using coded transmissions to serve multiple clients at once. While large gains in the rate of communication are obtained using coded caching, most existing schemes require that the files at the server be divisible into a large number of parts. In particular, most known coded caching schemes require subpacketization $ F={e^{Oleft ({{K^{frac {1}{r}}}}right)}} $, where K is the number of clients and r is some constant positive integer. While few schemes having subpacketization linear in K are known in literature, unfortunately such schemes require large number of users to exist or offer little gain in rate. In this work, we propose a class of coded caching schemes based on projective geometries over finite fields, generalizing recent results. Our construction achieves subexponential (in K) subpacketization, ie, $ F={q^{O …

Abstract

In this paper, we propose a new model predictive control (MPC) formulation for autonomous driving. The novelty of our MPC stems from the following results. Firstly, we adopt an alternating minimization approach wherein linear velocities and angular accelerations are alternately optimized. We show that in contrast to the joint optimization, the alternating minimization exploits the structure of the problem better, which in turn translates to reduction in computation time. Secondly, our MPC explicitly incorporates the time dependent non-linear actuator dynamics that captures the transient response of the vehicle for a given commanded velocity. This added complexity improves the predictive component of MPC resulting in improved margin of inter-vehicle distance during maneuvers like overtaking, lane-change, etc. Although, past works have also incorporated actuator dynamics within MPC, there has been very few attempts towards coupling actuator dynamics to collision avoidance constraints through the non-holonomic motion model of the vehicle and analyzing the resulting behavior. We use a high fidelity simulator to benchmark our actuator dynamics augmented MPC with other related approaches in terms of metrics like inter-vehicle distance, trajectory smoothness, and velocity overshoot.

Abstract

Coded caching was introduced as a technique of systematically exploiting locally available storage at the clients to increase the channel throughput via coded transmissions. Most known coded caching schemes in literature enable large gains in terms of the rate, however at the cost of subpacketization that is exponential in ( being the number of clients, some positive integer). Building upon recent prior work for coded caching design via line graphs and finite-field projective geometries, we present a new scheme in this work which achieves a subexponential (in ) subpacketization of and rate , for large , and the cached fraction being upper bounded by a constant (for some prime power and constant ) . Apart from this asymptotic improvement, we show that through some numerical comparisons that our present scheme has much lower subpacketization than previous comparable …

Abstract

Robotic grippers have been designed for grasping a wide variety of objects. This paper presents a novel design of a flexible gripper suitable for gripping circular objects having variable curvatures and different textures. Two rubber belts form the gripper arms, which are used for gripping objects. These rubber belts are attached to DC motors. The motors are fitted in the gripper base. Magnets are attached on the other ends along with an interlocking mechanism. The primary actuation in the gripper is brought about by the Shape Memory Alloy (SMA) wire fitted along the inner side of the rubber belts. On energizing the SMA, the rubber belts bend along with the pre-programmed SMA and the magnets on the two belts come closer. Subsequently, locking is achieved. This forms a loop around the gripping object. The grip of the rubber belts around the object is further tightened by winding them around the shafts of the motors to which they are attached. This helps the gripper to firmly grasp objects of variable diameters. Gripping has been successfully tested on pipes, metal poles, trees with thin and thick stems.