The document proposes a supplier bot to deliver necessary supplies to COVID patients in care homes and clinics with small staff. The bot would use image processing and color detection to navigate rooms based on supply requests entered by patients into a Google Sheet. It would initially follow a blue path marked by the creator's lanyard and enter rooms based on detecting different colored notebooks outside to deliver requested supplies.
The document proposes a supplier bot to deliver necessary supplies to COVID patients in care homes and clinics with small staff. The bot would use image processing and color detection to navigate rooms based on supply requests entered by patients into a Google Sheet. It would initially follow a blue path marked by the creator's lanyard and enter rooms based on detecting different colored notebooks outside to deliver requested supplies.
The document proposes a supplier bot to deliver necessary supplies to COVID patients in care homes and clinics with small staff. The bot would use image processing and color detection to navigate rooms based on supply requests entered by patients into a Google Sheet. It would initially follow a blue path marked by the creator's lanyard and enter rooms based on detecting different colored notebooks outside to deliver requested supplies.
Problem Statement 2: Day to day struggles of a covid affected patient
Idea Outline With hospitals filling up and treatment costs rising, many patients are instead being looked after at covid care homes or smaller clinics. But issues like having a small staff count and the risks of entering a covid patient’s room/ward makes these places a less viable option.
The idea of the supplier bot is to take necessary supplies to the
patients when called upon, allowing minimal interaction, and less dependency of the patients on the staff. Implementation I have thought of using image processing using OpenCV in python to implement this idea. The bot would follow a path, which would be marked using a particular colour, and outside each ward, there would be a patch of a different colour. Based on which patient ordered the necessary supplies, the bot will look out for the corresponding colour patch, and when the bot encounters it, it will enter the ward. Implementation (Process) • The patient checks the checkboxes of the supplies or assistance they need on their phones. • The corresponding google sheet cells change their values from “no” to “yes”. • The python code, using the google sheets API, notes the changes, and using ‘pyserial’, communicates with the Arduino and the bot starts following the track till the camera connected to the bot senses the corresponding colour, after which the bot assists the patient. • The corresponding google sheet cells values are changed from “yes” to “no” by the code. Implementation Method (Colour patch detection) To represent the track, I have used my ID-Card’s lanyard, which is blue in colour. And for the colour patches on the floor, I have used notebooks of different colours. I found the appropriate HSV limit values using a trackbar mechanism. Implementation Method (Path following) Initially, my implementation was based upon finding just the centroid of the contour of the path, and comparing that with the centre of the frame, and deciding whether to move left or right based on that. But to increase performance, I have included angle calculation, which allows the path following module to be very accurate and smooth. Code file and demo video This is the demo of the software implementation of the project. I randomly change the values of some cells from “no” to “yes” and upon running the code, we can see that the bot follows the track and when it encounters the particular colours, it responds accordingly, and all values of the google sheet cells are switched to “no” at the end since the bot completed all the tasks. Code link: https://drive.google.com/file/d/1fdw8B5KBu7CNCq1f3J4w4rtjjhmEyChn/view?usp =sharing JSON File: https://drive.google.com/file/d/1SZy4Q5oUFpbNK8P09dtX_q_emVWRJx- n/view?usp=sharing Video link: https://drive.google.com/file/d/1RbaoD5T_tVit0d_FkZbC80HCSsEXp9LN/view?usp =sharing Pyserial code and Arduino video During the implementation of the project on hardware, I faced many issues. initially I used an HC05 Bluetooth module and the bot responded as desired. But due to some failure in the module I had to shift to connecting the bot using the Arduino UART cable. Unfortunately, there was a sudden failure in the motor driver as well, although it worked for some time. The python and the Arduino files will be included, and along with that I have also shared a 10 second video of the bot when it was working before the issues popped up. Python code with pyserial: https://drive.google.com/file/d/1crwOt6ohI-AKv8fQ- H4W7eGI02X9FmYW/view?usp=sharing Arduino code: https://drive.google.com/file/d/18XYurpnHmYI-P3JYKxOlgXhMCiRbTwd-/view?usp=sharing Video: https://drive.google.com/file/d/1gcEGtKuGUMUq5JcUGyL1_n- WBQGDDhht/view?usp=sharing Key elements used • OpenCV module in python for image processing • Pyserial module in python to communicate with Arduino • Google sheets API • Arduino and chassis for movement
For my prototype, I have used an HC-05 Bluetooth module and the
video feed is from my laptop’s webcam, although, for complete implementation this can be developed using a Raspberry – Pi and a camera module. Potential Although this idea is focused upon helping covid patients get supplies, this simple skeleton could be used for many applications. For example, it could act like a waiter bot at a restaurant to deliver food, or in factories to carry heavy goods.
With further developments, such as using ROS and introducing SLAM,
bots like these could be used extensively for wider range of applications. Thank You