Mechanical, Electrical, and Optical Design
Physical Assembly
The physically assembled sensor with slightly extended form factor but identical electronics and optical design. The fully assembled sensor weighs less than 1kg and measures at a sensitivity less than 1 parts-per-billion, making it easily deployable on mid-sized commercial UAVs.
Upgrade From Previous Design
Two versions of the sensor, the open cell version Phantom II (right) and close cell version Chameleon (left) sitting on top of the previous NRC methane sensor. The weight was reduced from 25KG to 3KG for the closed cell sensor with no sacrifice in sensitivity due to highly custom electronics and embedded software.
Deployment
The device has been deployed across Canada including near oil & gas facilities in Northern Alberta and the Arctic coast in collaboration with Geological Survey Canada and Environmental & Climate Change Canada to quantify total methane emissions in the country. The sensor seen here is mounted on a DJI Matrice 300 drone as it is prepared to take measurements.
Results
This is the result of a single collection done by the sensor mounted on an UAV measuring methane emissions in the Mackenzie Delta region of Northern Canada. The heatmap is the result of a fusion between GPS data, ambient conditions, and concentration measurement. Higher concentrations are shown in green around the slump site.
Associated Publications
M. Beattie, C. Sun, R. MacLeod, N. Sabourin, P. Morse, J. Corbin, J. Oliaee, “Ultra-Lightweight Mid-IR Methane Sensor for UAV-based measurements”, Journal of Optical Express, Submitted for review
J. Oliaee, R. MacLeod, C. Sun, M. Beattie, S. Dallimore, P. Morse, “Measurement of natural methane emissions in the western Canadian arctic using an ultra-lightweight laser spectroscopic sensor aboard a small UAV”, CabCH4 symposium, Ottawa, Canada, May 2024
C. Sun, J. Oliaee, N. Sabourin, G. Smallwood, “Development of a Lightweight UAV-Mountable Open-Cell Methane Sensor Using Mid-Infrared Tunable Diode Laser Interfaced with a Low-Cost Embedded Platform.” Onboard Sensing, Analysis, and Reporting (OSAR) conference. Riverside, CA, March 2023
]]>
The soft exoskeleton works by having 2 motors mounted on the thighs to reel in the paracord of the exoskeleton. The cord travels down to the knee and back up, forming a double pulley. It then bypasses the thigh using soft PTFE tubes and attaches at the back, where strain gauges measure the tension in the cord and closes the control loop. When the cord is reeled in, the exoskeleton exerts positive extension work on the 3 joints most involved in lifting motion (knee, hip, and back), and therefore reducing muscle activation and metabolic costs of the user thus reducing the likelihood of injury.
Since the exoskeleton is actuated by soft paracord, the controller must be adapted to the reality that the actuators can only pull and not push, and that force can only be exerted in 1 direction. The controller hierarchy consists of 3 levels:
The Sixth Sense IMU is built on an ESP32S3 Wroom Mini microcontroller on a custom PCB. The two core sensors are:
Here, a previous version of the Sixth Sense is mounted along with the Microstrain 3DM-GX5 (Ultra high end IMU costing $2000+) on a test subject along with motion capture dots to track ground truth joint kinematics. The test results are then fed into the EPIC Lab’s machine learning algorithms that predict joint moments from kinematic input.
The comparison results in a R-squared of 0.99+, showing near identical output between the two sensors. This allows the lab to replace $2000+ IMUs with custom PCBs with a unit cost of $16 with no sacrifice in performance.
The sonic wind sensor features 3 orthogonal pairs of opposing sonar transceivers where the transmitters are periodically pulsed and the travel time to the receivers are measured. This time is equal to the distance between the transceivers divided by the sum of the speed of sound and the parallel component of the wind. Accurately measuring this time means 3 orthogonal components of wind speed can be calculated.
The Beaglebone black SBC is used since it can handle ultra-fast real time operation with assembly. A custom capeboard is used to amplify high-impedance digital signals from the BBB to +-10V low impedance output to power the piezo-electric speakers. The signals from receivers are then read as digital inputs by the BBB with constant polling to precisely calculate time required for the sound to travel. The circuit, along with the high-speed assembly code, allows the sensor to measure 3D wind at 1000Hz, compared to a traditional wind cup sensor which can only measure 2D wind at ~1Hz.
Here the ultrasonic sensor is deployed along with a traditional wind cup sensor to measure the long term stability of the sensor.
The Photothermal Interferometer (PTI) is used to measure the presence and morphology of black carbon particles as they are significant contributors to climate change world wide. The PTI works by overlaying two laser in the path of flow of carbon particulates, with one laser exciting the particles and the other measuring post-excitation emissions. The challenge for the flow channel I designed is therefore exactly overlapping all 3 elements to enable these measurements.
To overlap all elements, I designed a U-shaped channel for air flow carrying carbon particles with a window at the bottom sealed by a dichroic mirror. This dichroic allows the 632nm probe laser (red) to pass through while reflecting the 760nm heating laser (orange) directly into the same path and exits on the other side. The heating laser is injected at Brewster’s angle to minimize reflection and maximize power.
To validate the design in simulation, I performed computational fluid dynamics simulations with ANSYS Fluent to see the path of travel of particles seeded into the stream. Since the design required that the particles stay on the same path of the laser, I ensured that the flow across the channel was maximally laminar through modifying the cross-section geometry of the channel.
The model of the UAV is based on the Skywalker X-8 Aerodynamic profile modified to be 3D printable. The challenge is that the airframe is very large (a wing span of 2 meters) and therefore parts had to be printed in segments and assembled. The CAD work primarily consists of using surface lofts to match the shape of the X-8 airfoil while creating 3D printable internal spars and ribs.
Another major design challenge was the requirement for vertical takeoff and landing (VTOL) capabilities. We incorporated carbon rods which attaches to actuators on both ends of the wing, mounted on servos which allows the rotors to tilt during flight.
The Wall-E Jr robot is designed with principles of sustainable engineering in mind, which means it uses primarily recyclable materials (aluminum and rubber) and is simple to produce, assemble, and repair. It is designed to automatically drive around the streets of Ontario and collect and sort trash into four categories.
