Author: igf500

For the week before the break, no progress was made on my work with the sensors and the design report turned out quite poorly and unsatisfactory. As a result, I am behind the projected schedule. Therefore, next week will be critical and significant work must be done in order to catch up and deliver on our product.

The first order of business, will be to pick up our sensors from the ECE equipment desk and start the local testing right away so we can have metrics and data to feed to our NodeMcu as well as understand the behavior of the sensor in order to fine tune it with our microcontroller. Once we have the sensor up and running, we will have a few main variables that will guide our experimentation. These are: distance from the sensor, fabrics/textiles and colors used for detection as well as the angle from the sensor (both vertically and horizontally) at which we place our subject.

In addition, the data gathered from these trials will also inform our power consumption estimates and based on this we will modify the plans we outlined on the design report or simply go ahead approve the batteries we intended to use as appropriate and proceed to order them.

However, prior to starting the testing at the beginning of the week, I must carefully review the design specs of the sensors and draw sketches of simple electrical circuits depicting the configuration that will be used for testing as well as the one that involves our NodeMcu. Careful consideration must be taken when connecting these components since we do not want to damage or ruin the equipment we have. After the trials are completed, I will work closely with Nat to integrate the sensor and the NodeMcu in order to have a working module that can then be scaled to the quantity we are aiming for.

Once we have the sensor and NodeMcu running, as part of my contribution for this subsystem, my focus will shift to the mounting mechanism that will be used to place our sensor module on the selected gym equipment. As a result,  I will have to brush up on my knowledge of fabrication techniques and simple prototyping. For the encasing that will house our module I must analyze which method of fabrication like 3D printing, manual woodwork, among others, is best . If we decide to go the 3D printing route, this will require familiarity with CAD tools and other similar suites available in the Makerspace, as well as careful scheduling and planning to reduce the time wasted while we fabricate encasing prototypes. If we deem this to be inefficient or too complicated for the purposes of our project, then I will explore simpler and perhaps even cruder methods of fabrication.  However, we must keep in mind our requirements of stability and robustness to vibration of our module when making this decision. Thus, we will most likely choose a fabrication method and tools that provide a balance between sophistication and ease of implementation and use. We want our module to be strong and rigid but we also want a tool that is simple to use given that we need to make the most out of the time we have and we need to catch up and keep up with the schedule.

After this week’s design review presentation we received some feedback pertaining the initial design of our system. With the questions we were asked in mind, my tasks for this week consisted in further researching the components for our Sensor Module. One question that came up was the detection angle of our sensors. In all honesty, I was more focused on the range (distance of detection) of these sensors and I admit I did not factor into my calculations how the detection angle of our sensor affected our system. Therefore, I looked into the documentation of our sensor and investigated other alternative sensors just in case. After some  hesitation I believe that our current chosen sensor works well based on our implementation.

Having said this, professor Mukherjee brought up a possible different implementation of our solution. He discussed a multiple sensor approach to obtain more accurate occupancy measurements that were impervious to the external environment around the gym equipment (people, movement, etc.). Nonetheless, after some consideration I believe it is best to stick with our one sensor approach given the simplicity of the communication between the sensor and the NodeMcu. Our sensor is analog and the NodeMcu ESP8266 only has a single analog input which would make the incorporation of multiple of our selected IR sensors practically impossible. We would have to research other sensors primarily digital ones that can be incorporated seamlessly with the NodeMcu and come up with an algorithm to coordinate the multiple sensors to reach a decision on occupancy. This task does not seem so difficult. Yet other considerations were also discussed.

A point that came up is how invasive our system would be if we were to incorporate say three sensors on a treadmill. This is certainly feasible but perhaps the wiring of these sensors might result cumbersome for users. In addition we would have to factor in power consumption which would most likely not meet or battery life requirement. As a result, and in tandem with our size requirements, we believe that it is best to stick with only one sensor for now and experiment with calibrating and tuning our microcontroller with our sensor.

Due to this deliberation in implementation and sensor type, no orders were placed this week, which I anticipate will put me behind schedule. Nonetheless, this setback does not hurt the overall pace of the team given how we decided to divide the tasks and how compartmentalized our testing plan and schedule is. Our verification plan has a lot of small testing phases which will either way take up most of the time after the parts arrive, hence setbacks can be easily recovered from. In my case I will have to get started as soon as I can with a bit of an accelerated pace on the IR testing in order to meet the goals of integration we have.

As a result my plan for next week is to complete the order of parts and hopefully receive them as soon as possible to start experimenting and testing locally according to our testing plan. In the meantime, I will most likely take more concrete physical measurements of the gym equipment in order to analyze these as well as compare them with our equipment’s manuals and documentation. These metrics will get the ball rolling for the work that needs to be done with respect to the Sensor Module mounting mechanism and will aid in setting up our testing environment which will emulate the machines we will be working with. As a result, next week I will be checking out the Makerspace to investigate the resources available to fabricate our setup (wood, tubes, tools, etc.). I will also research CAD and other suites since we have decided to fabricate our encasing for the sensor module locally. From this investigation, I will decide how to proceed with the mount design. All of this will aid in developing our Design Report which is the deliverable for next week.

My task for this week consisted in developing ideas pertaining to the physical aspects of our system. So, this week I spent the majority of my time researching components and parts we would need to complete our project as well as ways to link them all together. We have decided to use IR sensors and I have proposed the use of either the Sharp GP2Y0A02YK or the Sharp GP2Y0A21YK0F. Both these sensors work much in the same way however their distance detection ranges differ with the former having more range than the latter. Theses sensors were considered in the first place given that they are quite easy to work with and simple to interface with other pieces of hardware we are considering.

From this research, we have decided that the hardware module attached to the gym equipment will consist of our IR sensor connected to a Nodemcu and a power supply to power both of these. We have opted for Nodemcu due to the fact that it is a low-cost and open-source IoT platform that’s based on other components we thought of initially like the ESP8266 chip. This module will be interfaced with a Raspberry Pi which will handle the IoT component of our design.

In addition, I surveyed the University Center’s gym to get a feel of the machines and equipment we will be basing our design on. We will be using treadmills, stationary bikes and stair climbers as our base. I noted the models and styles of the machines and based on these observations I believe it would be best to attach our physical module to a mount that works like a phone bike mount does. Many of these machines have a lot of tubes to which we could secure or module. Therefore an adjustable design with rigidity in mind is optimal to ensure our sensor is always properly positioned. Below is a picture of a possible template for our design of the physical encasement of our hardware module.

For next week I am aiming to polish the details of these physical modules before placing any orders. I need to consider the power consumption and other physical aspects of our module like weight among others to strengthen our design. Based on this research, I will consult with my teammates and decide which process would be best to fabricate our modules (3D printing, carving, or off the shelf solutions, etc.) In addition I want to have the communication protocols of our modules clearly planned out before moving forward with our orders.

Considering that our design is quite new compared to other groups, perhaps we might be a bit behind in terms of schedule. Nonetheless, based on what our team has discussed, this week’s progress seems satisfactory and we will work to keep the pace for next week. We understand that our process since the beginning of the course has been rocky and constantly shifting so much emphasis must be placed in terms of progress for the upcoming weeks.

Given that our team has decided to pivot and change our project entirely, the extent of my progress for this week encompasses researching technologies and concepts pertaining to our new idea. This new idea centers around gym occupancy, more specifically gym machines. We will most likely base the occupancy on motion and proximity; hence I have been looking into different types of sensors that we could use to accomplish the task of detecting a user/occupant. I’ve looked into ultrasound, IR, and LIDAR sensors as possible options for our implementation. At the moment, based on my findings I would suggest we use ultrasound, IR or a combination of both of these given that the task at hand for these sensors does not seem to require expensive components nor sophisticated measuring capabilities. Further investigation needs to be done to determine how these sensors will communicate and what the limitations and specs of these are. Based on that we can go ahead and order specific brands that are best suited for our task based in the information gathered. I’ve also looked into the communication between Wi-Fi cards and Raspberry Pi, yet I need to figure out the specifics of how the sensor will interact with the aforementioned components.