Updated: Jun 8, 2018
This was a project I undertook inspired by work allowed me to explore how a robot could navigate around a house and detect things in its environment.
The initial aim was to make them airborne as drones but it just wasn’t feasible until I had ensured I could navigate indoors accurately. in fact I made 3 robots in total but never really solved the problem of indoor navigation. I managed to get sensors of all kinds working, but once on the move the robots struggled to determine their position in the house or travel around autonomously. The great thing the last year gave me is a true passion to learn more about what’s possible in robotics. It has turned into a real hobby now where I spend as much time as I can with research and experiments.
So where do we go now with this?
Navigation has been the main challenge so I have decided for this year I need to create a low powered network not using wifi to allow my robots to talk to each other and to a central computer, also this network made up of many sensors and microprocessor’s will act as beacons, and allowing all the parts talk to each other
Its more about creating the sensor network (hardware) and programming the sensors locations to provide key information about the location and what’s going on there, such as are any people present or a passing robot, or what’s the temperature. Using signal strength may be a way to determine triangulation or id signals using infrared or RFID from the sensor arrays, this will need exploring in full and a solution being created.
Initial plan for navigation network
This plan would Include:
Indoor Sensor network that can also be used to assist in robot navigation and also provide situation and environmental information.
Sensors and network should use low power, so can run on batteries for extended periods
Has to allow sensors to relay data back to base station or passing robots.
Should be able to transmit location information
Should allow visual monitoring from central computer or location using Web browser.
To achieve this I will use a combination of:
5 – 10 Arduino Nano’s (Wireless enabled using NRF24L01’s)
3 x Raspberry Pi’s (On robots and network control)
5 – 10 NRF24L01 Transceivers
Rechargeable power supplies
IR Transmitter and Receiver LEDs ( for robot to robot identification and location info)
Ultrasonic short range detectors (up to 4 Meters)
LIDAR Mid to long range sensor (on main security robot)
PIR – Detectors (human and lifeform detection)
Temperature and Sound detectors
Potentially use RFID for close proximity
The key here is to do this with really cheap components and not off the commercial finished products, so network transceivers costing just a couple of $ each.
I have about half the equipment required already and would use the 3P monies to add some of the missing elements.
Success Criteria I was looking for
There are varying degrees of success on this, but the main one for me would be having a successful homemade network with its own protocol/language to talk to the sensors and robots, allowing the robots to determine more accurately where they are.
The system should allow monitoring remotely and online ideally so I can check how things are even when not at home.
I managed to prove all of the concepts of the navigation methods I set out to achieve. This will be a continued area in my research and creation of automated Robots.