Time is running and I am away from home with my kits I don't add additional hardware functionality yet but change the user interface from the previous one. I planned to add RSSI measurements but Ground Operations Centre took all the time. This article is the final one for this design challenge therefore I will explain what I have done till now and what can be added in the future.
Previously, the helmet was sending one data per second. Now, it is sending 100 data pre second. I accomplish this by decreasing the system tick from 100ms to 10ms and task pause from 10 ticks to single tick. 10 times from the system speed and 10 times from the waiting overall data transmission is 100 folded. It can be increased more but it requires a comprehensive timing analysis which can be a future task.
Hence, I am not in my town I just grab the MSP432 with sensors and leave the helmet and gate control circuit at home. This is not my first trip during the design challenge and kits travelled Canada before they have arrived so they have seen many places
Let's see how the system is working now.
The following video is the final hardware version for the design challenge. Unfortunately, I couldn't carry all parts with me. That is why demonstration above only show acceleration and temperature. You can find the details here Safe & Sound Wearables - Trackable Safety Helmet for Miners #12: Smart Helmet Video Demonstration .
I implemented two graphs; one for temperature and one for acceleration. There are also alarm indicators and alarm clear button. The controller can see 10 seconds window graph but also able to see the past data by dragging the cursor.
Initial plan and how things have changed
My idea aims to improve the mine safety. It will track the miners, detect possible explosions, and prevent human ignorance while forcing them to wear proper work suit. I planned the following tasks before I started the project. I achieved some, some are failed, and some are impossible to implement with current tech.
1 - As I mentioned above, the main problem is high methane gas concentration. If the air includes 5% to 15% methane it can explode. Because of this, methane concentration must be observed. This should be placed in many areas of the mining area but if we implement the sensor on the helmet, we can see how the concentration change around the workers. How concentration change in a time and place will give important information about the estimating explosions. Methane gas sensors, like Methane CNG Gas Sensor - MQ-4, usually have analogue output and MSP432P401R has 14-bit 1MSPS differential SAR ADC. The first task is connecting an analogue gas sensor to MSP432P401R and process it.
2 - Methane concentration is the obvious indicator of the explosion. However, there are other important parameters like pressure and temperature so they should be measured. BOOSTXL-SENSORS booster pack has an environmental sensor which can measure ambient temperature, pressure, and humidity. This information can be analysed and gas concentration limitation can be changed dynamically.
3 - In order to achieve a reliable system, it is required to have a communication link. The main reason is the collected data can be analysed by the central point and if something is wrong warning signal can be activated. It also brings other benefits like if something happens to the worker, he can send an emergency message. The kit includes Bluetooth and Wi-Fi network processors. Bluetooth is not suitable hence it is not supporting meshing (Bluetooth 5 bring this support) and distance is short. Wi-Fi is the better option because it has mesh support and longer range. However, it is still not enough distance, if we imagine the mine area and any explosion can disrupt communication if a link is destroyed. Long-range communication link will be feasible for this application like narrow band communication. The required data rate is not high so Lo-Ra, SigFox, or sub 1GHz long range RF can be preferred. Hence, the mine are under the ground and there may be lots of distractions during the path, this is the hardest part of the project. I will implement the project by Wi-Fi but it should be changed considering the outline of the mine. Implementation will be similar only the RF part and protocol is changing.
4 - Another problem is after an explosion it takes long times to reach the miners sometimes it takes months. If we can build a tracking system it will help us to reach them swiftly. Actually, the previous task also cover this topic. If we implement a reliable mesh network, we can use both time of flight (ToF) and received signal strength indicator(RSSI) to track the miners. However, inside the mine, both of these may not give the accurate results. This should be supported by the dead-reckoning. Dead-reckoning can be implemented by using inertia measurement unit (IMU) which includes accelerometer and gyroscope. BOOSTXL-SENSORS booster pack also includes IMU sensor. Dead-reckoning can be synchronised at the mesh points so accuracy can be increased.
5 - Hence, we have IMU sensor for tracking, we can add additional functionality. We can detect the falling of the workers and help them immediately. This looks unnecessary because of the emergency message but if the miner loses the conscious, the automated fall detection can help to find worker immediately.
6 - Some people are ignorant and unless you do not push them to wear safety uniform, they don't wear it. In order to be sure helmet is worn, workers should be controlled at the mine access area and helmetless entry shouldn't be allowed. To achieve this RFID-enabled ticket can be placed inside the helmet and gates are opened by the helmet. Bluetooth can also be preferred if the power consumption problem is solved for wearable. With the implementation of the Bluetooth, the helmet can act as a Bluetooth smart key fob and access is denied without the helmet.
Challenges and Threats
The hardest part of the project is designing a reliable communication network. It is one of the key parts of the project. If a reliable network connection is provided, it will be so easy to implement emergency messaging and make the tracking easier. There are lots of computationally intensive task for the processor to do, I am not sure one processor can handle all of these so multiple processors may be required but this will increase price and size. It is also possible to send the data to the centre and make calculation there and the send back to the result like the idea in the IoT systems. Dead-reckoning can be also challenging if the proper synchronisation and filtering are not done. Processing data (correlation between different variables) will require a comprehensive research for implementing this system into real mine. I may implement an estimated or dummy correlation hence the real data correlation includes different disciplines and measurements inside the real mine.
The main task was measuring the methane concentration which was failed. Before applying for the design challenge, I just heard about the ATEX but didn't know much about it. I did some research, get comments/help from the Element14 community and it looks like impossible to implement because I couldn't find a small gas sensor which can fit inside a helmet and ATEX approved. The problem gas sensors require a heating element so they require an ex-proof box or could spark an explosion. This was my fault in the planning part and there was nothing much to do. I do implement a temperature sensor and a pressure sensor. Of course, they are not intrinsically safe yet but it is possible to implement.
The communication was another important task. The TI https://www.element14.com/community/view-product.jspa?fsku=2535895&nsku=38Y6657&COM=noscriptSimpleLink Wi-Fi CC3100 module BoosterPack works fine but I test it inside a room. I am sure mines has lots of things to block or interfere the Wi-Fi signals. Therefore, it requires testing inside the mine area but unfortunately, I don't have a chance to go inside a mine even I could the design should be approved as intrinsically safe before the tests.
The next task was tracking and detecting the faints. I have used an ADXL345 accelerometer which can detect the inactivity and free-fall. I successfully detect and transmit the free-fall and inactivity. I haven't had time to implement dead-reckoning. It is not easy to implement an accurate dead-reckoning. It can be a future work to add an IMU sensor and implement a dead-reckoning with proper filters. RSSI signals also can be used to improve the tracking and decrease the deviation.
The final task was to prevent the human ignorance. I think this task has a pretty solid success. If RFID tag is inserted inside the helmet during the manufacturing miners has no choice to access without the helmet. It can be also used for tracking the miners and workload. Hence, a passive tag has inserted the helmet, it doesn't require ATEX certification.
I can divide the software into two parts. One firmware for the MCUs and the other one is the PC program Ground Operation Centre. For MCU programs, I prefer to use TI-RTOS. I hadn't used RTOS before the design challenge. It was a little difficult at first but I think it makes things easier for me for the later changes like increasing the data transmission. My idea was also challenging me and improve myself. I think it was a good intro.
The Ground Operation Centre is the backbone of the project. It collects data and shows it a user-friendly interface. I think it requires a database later on for recording all the data for a longer time. It can be integrated into work process and simplifies the things while improving the safety. I have coded the program in C#.
You can download the whole project from the following links.
Ground Operation Centre : https://drive.google.com/open?id=0B9wed2dhooCNTDJRZVBIV0p3YjQ
The main hardware component of the design is MSP-EXP432P401R LaunchPad. It does pretty much everything. I also used the following components in my project.
The Wi-Fi boosterpack to send data to Ground Operations Centre Texas Instruments: SimpleLink Wi-Fi CC3100 module BoosterPack
Accelerometer Sensor Analog Devices: ADXL345
Temperature Sensor: Texas Instruments: TMP102
Force Sensitive Resistor: Interlink Electronics: FSR402
Dead-reckoning is the main thing I should implement the future version. To improve accuracy, RSSI data will be useful. The design should include some hardware implementations like using fuses and Zenner diodes. It is also required that power analysis of the circuit which is required for the ATEX and also for the battery life.
It was honouring and useful for me to take in part in this design challenge. I would like to thank Element14 community for their help. I also want to thank Element14 staff and TI for organising the challenge and providing us this opportunity.
You can see all the links related to this project in the first blog: Safe & Sound Wearables - Trackable Safety Helmet for Miners #1: Introduction to Project