The traditional understanding of robotics includes mechanisms of rigid structures, which can manipulate surrounding objects, taking advantage of mechanical actuators such as motors and servomechanisms. Although these methods provide the underlying fundamental concepts behind much of modern technological infrastructure, in fields such as manufacturing, automation, and biomedical application, the robotic structures formed by rigid axels on mechanical actuators lack the delicate differential sensors and actuators associated with known biological systems. The rigid structures of traditional robotics also inhibit the use of simple mechanisms in congested and/or fragile environments. By observing a variety of biological systems, it is shown that nature models its structures over millions of years of evolution into a combination of soft structures and rigid skeletal interior supports. Through technological bio-inspired designs, researchers hope to mimic some of the complex behaviors of biological mechanisms using pneumatic actuators coupled with highly compliant materials that exhibit relatively large reversible elastic strain. This paper begins the brief history of soft robotics, the various classifications of pneumatic fluid systems, the associated difficulties that arise with the unpredictable nature of fluid reactions, the methods of pneumatic actuators in use today, the current industrial applications of soft robotics, and focuses in large on the construction of a universally adaptable soft robotic gripper and material application tool. The central objective of this experiment is to compatibly pair traditional rigid robotics with the emerging technologies of sort robotic actuators. This will be done by combining a traditional rigid robotic arm with a soft robotic manipulator bladder for the purposes of object manipulation and excavation of extreme environments.

In order to refine autonomous exploratory movement planning schemes, an approach must be developed that accounts for valuable information other than that gained from map filling. To this end, the goal of this thesis is divided into two parts. The first is to develop a technique for categorizing objects detected by an autonomous exploratory robot and assigning them a score based on their interest value. The second is an attempt to develop a method of integrating this technique into a navigation algorithm in order to refine the movements of a robot or robots to maximize the efficiency of information gain. The intention of both of these components is to provide a method of refining the navigation scheme applied to autonomous exploring robots and maximize the amount of information they can gather in deployments where they face significant resource or functionality constraints. To this end this project is divided into two main sections: a shape-matching technique and a simulation in in which to implement this technique. The first section was accomplished by combining concepts from information theory, principal component analysis, and the eigenfaces algorithm to create an effective matching technique. The second was created with inspiration from existing navigation algorithms. Once these components were determined to be functional, a testing regime was applied to determine their capabilities. The testing regime was also divided into two parts. The tests applied to the matching technique were first to demonstrate that it functions under ideal conditions. After testing was conducted under ideal conditions, the technique was tested under non-ideal conditions. Additional tests were run to determine how the system responded to changes in the coefficients and equations that govern its operation. Similarly, the simulation component was initially tested under normal conditions to determine the base effectiveness of the approach. After these tests were conducted, alternative conditions were tested to evaluate the effects of modifying the implementation technique. The results of these tests indicated a few things. The first series of tests confirmed that the matching technique functions as expected under ideal conditions. The second series of tests determined that the matching element is effective for a reasonable range of variations and non-ideal conditions. The third series of tests showed that changing the functional coefficients of the matching technique can help tune the technique to different conditions. The fourth series of tests demonstrated that the basic concept of the implementation technique makes sense. The final series of tests demonstrated that modifying the implementation method is at least somewhat effective and that modifications to it can be used to specifically tailor the implementation to a method. Overall the results indicate that the stated goals of the project were accomplished successfully.

The presented work in this report is about Real time Estimation of wind and analyzing current wind correction algorithm in commercial off the shelf Autopilot board. The open source ArduPilot Mega 2.5 (APM 2.5) board manufactured by 3D Robotics is used. Currently there is lot of development being done in the field of Unmanned Aerial Systems (UAVs), various aerial platforms and corresponding; autonomous systems for them. This technology has advanced to such a stage that UAVs can be used for specific designed missions and deployed with reliability. But in some areas like missions requiring high maneuverability with greater efficiency is still under research area. This would help in increasing reliability and augmenting range of UAVs significantly. One of the problems addressed through this thesis work is, current autopilot systems have algorithm that handles wind by attitude correction with appropriate Crab angle. But the real time wind vector (direction) and its calculated velocity is based on geometrical and algebraic transformation between ground speed and air speed vectors. This method of wind estimation and prediction, many a times leads to inaccuracy in attitude correction. The same has been proved in the following report with simulation and actual field testing. In later part, new ways to tackle while flying windy conditions have been proposed.