Pratap Bhanu Solanki

LED-based underwater wireless optical communication is emerging as a promising, low-to-medium range alternative or complementary technology for acoustic communication, because of its attributes of low power consumption and high data rates. This work demonstrates a use case of the optical communication technology, where an underwater robot is controlled via blu-light optical signal that is transmitted by a human operated joystick. For implementation details and background, refer to the this paper.

Achieving and maintaining line-of-sight (LOS) is challenging for underwater optical communication systems, especially when the underlying platforms are mobile. In this work, we propose and demonstrate an active alignment control-based LED-communication system that uses the DC value of the communication signal as feedback for LOS maintenance. The effectiveness of the approach is first evaluated in simulation by comparison with extremum-seeking control, where the proposed approach shows a significant advantage in the convergence speed. The efficacy is further demonstrated experimentally, where an underwater robot is controlled by a joystick via LED communication. For detailed background in alignment control refer to the this paper.

Networks based on optical communication have dynamic links. These links only exists when there is a close to Line-of-Sight (LOS)between the corresponding robot nodes. In this work we first formulated a scenario of multiple robots equipped with optical communication devices. Randomness is introduced in the rotational behavior of optical device to increase the chances of getting (LOS). For simulation purpose, we defined some protocols which can be practically realized to start and maintain the communication link. Simulation results shows the feasibility for convergence of agreement protocol in such a system. For instance, the implementation of formation control is illustrated in the video. The details are included in this report.

Muscle activity at the arm are responsible for enabling many human tasks. We developed a wearable device using force-sensing resistors for detecting the inflation of different muscle groups at the forearm and upper arm. By capturing the pressure generated at the muscle belly, the sensing straps were used as a human interface device for controlling the computer keyboard and mouse cursor. In addition, we implemented a neural network for recognizing the patterns produced by multiple arm and hand gestures. Finally, force-sensing resistor shows promising application since they are unobtrusive, lightweight and has fast response. For more details, please refer to this report.

Fish-like swimming could have considerable amount of energy savings while it is used for small autonomous underwater vehicle. But, analysis of fish motion dynamics in the water, demands great attention. Here, in this work, we analyze fish motion dynamics in the water and design a robotic fish which is capable of swimming in water. The skeleton of the fish which consists of 3-joint mechanical tail (rear part) and two pectoral fins, are designed and fabricated. Experiment shows that the mechanical tail is capable of generating near sinusoidal wave through its body which is essential for the fish-like swimming behavior of the robot. For more details, please refer to this paper.

Robot Motion planning is one of the fundamental problem in robotics. It has various applications such as in Ware house automations, Driverless cars, Robotic surgery etc. The path planning problem is described as: to find a shortest or optimized path between start point and goal in a spatial configuration consists of obstacles of various types. In this project we extended the Voronoi partitioning from point obstacles to polygonal obstacles. Additionally, we used Djikstra's graph search algorithm to compute the shortest collision free path from user defined initial and desired final configuration. For more details please refer the report and the project's Github page.

This project aims towards calibrating the images projected on arbitrary surfaces. We use projector and camera system to calibrate the projected im- age. Calibration is done by first recognizing the geometric shape of the projected surface. This is done by projecting binary coded images on the surface and then calculating transformation in various parts of the images. Now we pre-distort the new image before projecting on the surface in order to compensate for the distortion done by the surface. This method is applicable on any kind of static arbi- trary surface. Please refer to the project's report for more details.

In this project we made a metal prototype of famous commercially available balancing bird toy, a bird which balances itself on its beak . First we made a thermocol prototype by thermocol carving. We tried to balance this on its beak by tweaking the shape of thermocole by carving. Later using lost foam casting we got an iron prototype. It did not get balanced on the first run. We tweaked its mass distribution it by filing and grinding the bottom part. After few attempts it was able to balance on its beak as it can be seen from the video.