Abstract

A vision based exploration algorithm that invokes semantic cues for constructing a hybrid map of images - a combination of semantic and topological maps is presented in this paper. At the top level the map is a graph of semantic constructs. Each node in the graph is a semantic construct or label such as a room or a corridor, the edge represented by a transition region such as a doorway that links the two semantic constructs. Each semantic node embeds within it a topological graph that constitutes the map at the middle level. The topological graph is a set of nodes, each node representing an image of the higher semantic construct. At the low level the topological graph embeds metric values and relations, where each node embeds the pose of the robot from which the image was taken and any two nodes in the graph are related by a transformation consisting of a rotation and translation. The exploration algorithm explores a semantic construct completely before moving or branching onto a new construct. Within each semantic construct it uses a local feature based exploration algorithm that uses a combination of local and global decisions to decide the next best place to move. During the process of exploring a semantic construct it identifies transition regions that serve as gateways to move from that construct to another. The exploration is deemed complete when all transition regions are marked visited. Loop detection happens at transition regions and graph relaxation techniques are used to close loops when detected to obtain a consistent metric embedding of the robot poses. Semantic constructs are labeled using a visual bag of words (VBOW) representation with a probabilistic SVM classifier.

Abstract

An adaptive partition based Random Forests classifier for outdoor terrain classification is presented in this paper. The classifier is a combination of two underlying classifiers. One of which is a random forest learnt over bootstrapped or offline dataset, the second is another random forest that adapts to changes on the fly. Posterior probabilities of both the static and changing/online classifiers are fused to assign the eventual label for the online image data. The online classifier learns at frequent intervals of time through a sparse and stable set of tracked patches, which makes it lightweight and real-time friendly. The learning which is actuated at frequent intervals during the sojourn significantly improves the performance of the classifier vis-a-vis a scheme that only uses the classifier learnt offline or at bootstrap. The method is well suited and finds immediate applications for outdoor autonomous driving where the classifier needs to be updated frequently based on what shows up recently on the terrain and without largely deviating from those learnt at bootstrapping. The role of the partition based classifier to enhance the performance of a regular multi class classifier such as random forests and multi class SVMs is also summarized in this paper.

Abstract

In this paper a novel suspension mechanism for rough terrain mobility is proposed. The proposed mechanism is simpler than the existing suspension mechanism in the sense that the number of links and joints has been significantly reduced without compromising the climbing ability of the rover. We explore the use of compliant elements like springs for passively controlling the degree of freedom of the proposed mechanism and a framework for optimizing the spring parameters has been proposed. A performance evaluation of the proposed mechanism has been shown in terms of extensive simulations.

Abstract

In this paper we present two mechanisms of linear force actuator based actively articulated suspension vehicles and propose a strategy to control the wheel–ground contact forces to improve traction and to increase the no-slip margin and hence enhance the mobility of the vehicle on uneven terrain. We present the quasi-static analysis of each of the mechanisms to depict the ability of the systems to control the wheel–ground contact forces while negotiating uneven terrain with the help of feasibility plots. The first model is a vehicle with a 1-dof leg (referred to as LFA-V1) and can climb slopes upto 40 degrees but to further increase the capability of the robot we come with a modified design of the vehicle which has a 2-dof leg (referred to as LFA-V2) and can negotiate slopes with discontinuities greater than twice the wheel diameter.

Abstract

Detection of moving objects is a key component in mobile robotic perception and understanding of the environment. In this paper, we describe a realtime independent motion detection algorithm for this purpose. The method is robust and is capable of detecting difficult degenerate motions, where the moving objects is followed by a moving camera in the same direction. This robustness is attributed to the use of efficient geometric constraints and a probability framework which propagates the uncertainty in the system. The proposed independent motion detection framework integrates seamlessly with existing visual SLAM solutions. The system consists of multiple modules which are tightly coupled so that one module benefits from another. The integrated system can simultaneously detect multiple moving objects in realtime from a freely moving monocular camera.

Abstract

This paper depicts the utility of having robot formations to aid in collision avoidance amongst multiple robots in a multi agent/multi robotic setting. A set of robots satisfying certain proximal constraints reach a formation that enables considering that set of robots as a single robot cluster. Such a robot cluster is characterized by a cluster velocity computed from the individual robot velocities that comprise that cluster. The multi robotic collision avoidance problem can then be posed as collision avoidance between such robot clusters than between individual robots. Clusters could include single robot clusters. The robots within a cluster move such that they are guaranteed not to collide with each other. This is accomplished in this paper by ensuring that robots within a cluster reach a configuration where their velocities are equal and they move in a parallel formation. The underlying collision avoidance scheme thus needs to search for cluster velocities that provides collision free motion between clusters; this search at the cluster level than at an individual robot level greatly reduces the search space, thus providing for a hierarchical collision avoidance strategy. The strategy has been tested and verified in simulations and their results presented vindicating its efficacy.

Abstract

In this paper we describe the mechanical design, gait and control for a modular hexapod robot which makes it compliant to terrain while climbing slopes up and down. The robot uses a unique electronically actuated 2DOF universal spine similar to the spine of the Snake and the four legged animals like leopard, tiger etc. By controlling the amount of stiffness of the spine we were able to make the robot compliant to the surface. To our knowledge this is the first modular hexapod robot which achieves controllable compliance by both mechanical design and electronically actuating the robots spinal backbone. This design was biologically inspired based on the structure of a caterpillar with legs. Also we present the hardware and software architecture of the robot in this paper.

Abstract

This paper proposes a new framework for data association to solve the problem of SLAM. The proposed framework has specific relevance to range scanner based EKF-SLAM. The resulting data representation enables semantic reasoning on a spatial level which reduces the misassociation of closely spaced data from different spatial configurations through the use of convex polygons to represent data from similar spatial configurations. The data representation is especially effective for association when revisiting previously mapped regions efficiently. The spatial data representation also builds an occupancy grid for the entire map. We also provide a means of clustering range scan data using an adaptive threshold to be able to divide data at various ranges into clusters and dense data clustering to get more accurate data.

Abstract

Here in this paper we present a snake robot which can autonomously change its gait depending on the terrain. Also we present the design and construction details for the same. Here a custom simulator for the robot and an API for the same was developed which reduced the design and development time. For controlling the Snake robot manually a data glove using accelerometers was developed. The Snake robot can also be operated in autonomous mode where an over camera is used for sensing the terrain and the snake autonomously travels from one part of the terrain to another by switching its gait

Abstract

In this paper we describe the robot software development framework with database at its core. In this framework the features of database is harnessed to make various components modular, and robot data handling and processing is made much more flexible. This empowers the researchers to design the system quickly, access the data and allow multiple collaborators work to be integrated with ease. The advantage of database based framework is explained case by case for various scenarios with example robot application developments.