RESEARCH

Project Title: GPS Based Path Planning Algorithm for Agriculture Drones

Drone technology is advancing in rapid pace in the past decade, particularly in agriculture and its allied fields it has been extensively used. Using drones, made farming productive and precise. They help in optimizing agricultural operations, monitoring crops and their growth, and help in monitoring soil health with help of sensors and digital image processing. Piloting a drone is not easy and requires some level of expertise. So, a novice user cannot pilot a drone without proper training. To overcome this, an autonomous drone can be used. In any autonomous mobile robot path planning is an important subsystem. In this work, a GPS based path planning algorithm is proposed which can aid farmers in various agricultural activities. The system takes inputs related to farm boundary and crop planting pattern and generates path output based on the type of agriculture activities like surveying, spraying, seeding and quick monitoring. Cubic polynomial and spline function are used to calculate the path and its via points. The proposed algorithm is implemented in MATLAB.

Student: Sriram Reddy Gade

Official MARS Rover team from Amrita School of Engineering, Bengaluru campus.

Know More

Version-1

Agriculture is quickly becoming an exciting high-tech industry, drawing new professionals, new companies and new investors. The technology is developing rapidly, not only advancing the production capabilities of farmers but also advancing robotics and automation technology as we know it. Robots are equipped in the field of sowing, ploughing, harvesting and etc. making the agricultural process completely automatic. Main hurdle faced in automated harvesting is on the fly detection of produce and its location. Current state of art is ‘stop-and-go’ harvesting mechanism which will slow down the overall harvesting speed.

To do this task efficiently we need a quick light weight algorithm to identify the produce in a complex background and locate it accurately. The aim of this work is to develop an algorithm to detect the produce in a complex background and identity its location. An algorithm that would be equipped in an agricultural robot which would detect the most suitable produce for optimum profit, process where its stem is and allow the robot to harvest the vegetable from the plant on the fly. This algorithm would allow scanning, detection and selection of the right size of the cucumber for harvesting.

Students: Mr. Sachin C, Mr. Manasa N & Mr. Vicky Sharma

Version-2

In recent years, with the rise in deep learning frameworks, the number of fruit and vegetable detection systems also increased. These frameworks sup-port the automation of the harvesting process. With the harvesting process auto-mated, a significant amount of harvest and post-harvest waste can be prevented compared to when done manually. In this paper, we propose an economical and efficient algorithm to detect the produce and identify the stem coordinates of the appropriate produce to harvest by distinguishing between foreground and back-ground produce using stereo vision and finding the distance from the harvesting system to the produce. The motivation behind using stereo images is to facilitate the process mentioned above in real-time economically. The input will be a live video input from which a pair of stereo images are extracted at a particular sam-pling rate. Produce is detected using a modified version of the You Only Look Once algorithm, and a depth map is generated using triangulation to find the dis-tance and coordinates. The model's output will show bounding boxes around de-tected produce with 91.5% accuracy and finds the distance of stem coordinates with a high accuracy and a low RMS error of 0.13.

Students: Nimalan Karthik R, S. Manishankar, and Srikar Tondapu

Mobile robots are extensively used in service-based environments like warehouses, hotels, hospitals, restaurants etc. Path planning to reach destination from a source is a crucial task in any mobile robot application. An idle path planning algorithm should plan a shortest path with less computation time autonomously. Main aim of this work is to develop an ideal path planning algorithm for indoor mobile robots using QuickResponse(QR)codes as via points. To implement and test the algorithm in an in-house build mobile robot powered by Beagle Bone Black.

Students: Mr. Dhamodhar Reddy & Mr. Raviteja

The aim of this work is to design and develop a anthropomorphic arm to fold cloth. Our work aims in developing an algorithm for teaching a manipulator to perform a task. Teaching pendant simplifies the process of teaching for the end user and also trains manipulator to a user specific problem. In this work, an approach to teaching a manipulator using Learning from Demonstration technique is proposed.

Students: Mr. Bharadwaj Sannapaneni & Mr. Marut Shaswat

The aim of the project is to design and develop a six legged robot to navigate in an uneven surface. Irrespective of the surface evenness the robot will navigate and the chassis will be parallel to the flat surface.

Students: Mr. Sreejith, Mr. B Veekshan Sree Sesha Sai, Mr. B Akshay Kumar & Mr. B Mani Rajesh Reddy

Multi-robot system is gaining its importance in robotic research. One critical issue in multi-robot system is collision among the mobile robots while sharing same workspace. This paper deals with the collision-free path planning for multiple mobile robots using Bluetooth RSSI value. In the proposed collision avoidance algorithm a decentralized approach with fixed priority level for robots is considered. A variable speed technique based on the RSSI value of robots is used and the obstacle avoidance is implemented based on State based Obstacle Avoidance Algorithm. Proposed algorithm is implemented and tested using Webots 3D simulator.

Students: Ms. Lijina P

The paper describes the design of pick and place 3-axis SCARA robot with a compact universal gripper. Pick and place an object is the predominant job of the robotic manipulators in the industry. The challenging task in designing such manipulators is to develop a universal gripping mechanism, which should possess the ability to pick unfamiliar objects irrespective of its shape within the given size range. The most commonly used technique to accomplish such challenge is multi-fingered approach, but it increases the complexities on both hardware and software for smaller objects. The gripping mechanism discussed in this paper uses the concept of granular jamming. The gripper consists of single mass of granules which conforms to the shape of the target objects when pressed onto it. On evacuating the air, granules contracts and hardens quickly to hold the object. This approach of gripping mechanism does not require any sensory signal to pick up an object.

Students: Mr. Balaji A

Mobile robots are extensively used in service-based environments like warehouses, hotels, hospitals, restaurants etc. Path planning to reach destination from a source is a crucial task in any mobile robot application. An idle path planning algorithm should plan a shortest path with less computation time autonomously. Main aim of this work is to develop an ideal path planning algorithm for indoor mobile robots using QuickResponse(QR)codes as via points. To implement and test the algorithm in an in-house build mobile robot powered by Beagle Bone Black.

Students: Mr. P. R. Teja, Mr. Bhusapalli Dhamodar Reddy

Students: Mr. Aji S, Mr. Sri Harsha & Mr. Prashant Burnwal