You’re about to embark on an exciting journey with Stanley, a DIY NVIDIA Jetson Nano drone designed for in-flight AI research. This article provides an engaging look at the entire build process, showcasing the step-by-step assembly and calibration of Stanley. From the selection of key components—such as the PixHawk flight controller, Jetson Nano, and OAK-D Lite camera—to tackling the unique challenges of drone construction, you’re guided through each stage with enthusiasm and hands-on advice.
The creator shares fascinating lessons learned from past missteps, such as the meticulous installation of motors and propellers to avoid mishaps during flight. You’ll watch as Stanley takes shape, gaining insights into the technical nuances that enhance his abilities. With aspirations for future autonomous features using cutting-edge technology, this project also encourages you to dive into the world of DIY drones. You’re invited to engage with the content, build your own version, and embrace the trial-and-error process that makes drone creation so rewarding.
Meet Stanley, your new research companion for in-flight AI, equipped with an NVIDIA Jetson Nano and an OAK-D Lite.
Today, you’ll explore the COMPLETE build process, from individual components to a fully operational system. I’ve covered as many steps as possible to let you follow along and create your own!
Please feel free to ask any questions in the comments, and I’ll try my best to answer!
Part 1: • Your Raspberry Pi drone: the story so far
Part 2: • Stereo depth mapping with OpenCV and …
Part 3: • You Only Look Once: object detection …
Part 4: (this video)
Part 5: • Teaching your custom AI drone to track …
🌐 ESC Wiring Guide:
https://ardupilot.org/copter/docs/con…
⚙️ Parts Used
PixHawk 2.4.8
https://www.ebay.com/sch/i.html?_nkw=…
(Prices seem to be crazy for this board on e.g. Amazon, so make sure to shop around)
NVIDIA Jetson Nano (4GB variant)
https://www.amazon.co.uk/NVIDIA-Jetso…
S550 Frame
https://www.amazon.co.uk/FPVDrone-Hex…
Ublox 7M GPS Module
https://www.unmannedtechshop.co.uk/pr…
HobbyPower 2212-KV920 Motors (3x CW, 3x CCW)
https://www.amazon.co.uk/Hobbypower-9…
Battery (LiPo, 4S, 5000mAh, 25C – 50C)
https://hobbyking.com/en_us/turnigy-b…
ESCs (6x, SimonK, 30A)
https://www.aliexpress.com/item/32252…
F701 DSMX Reciever
https://www.aliexpress.com/item/32214…
Turnigy SBEC (5V / 6V, 5A)
https://hobbyking.com/en_us/turnigy-5…
Most parts were bought as a kit, so it might be easier for you to do the same.
🎞️ Chapters
😎 Social Stuff
Twitter: / akamatchstic
Patreon: / akamatchstic
ℹ️ Attributions
More than eight million trees lost in storms in the UK – BBC News
• More than eight million trees lost in…
🎵 Music
TABAL – An Unknown Journey
Provided by Lofi Records
Watch: • Video
Gigakoops – Spider Nest Castle
Listen: https://gigakoops.bandcamp.com/track/…
Alex-Productions – Back Home
Watch: • Travel Music by Alex-Productions (No…)
Alex-Productions – Nostalgia
Watch: • Motivational Travel Vlog Music by Ale…
Nihilore – Oblivious
Download: http://www.nihilore.com/synthwave
Corbyn Kites – Instant Crush
Watch: • corbyn kites – instant crush [synth…]
Vans in Japan – Deep State
Watch: • Video
Scott Buckley – Icarus
Watch: • ‘Icarus’ [Uplifting Epic Orchestral C…]
Scott Buckley – Undertow
Watch: • ‘Undertow’ [Sombre Piano \u0026 Strings CC…]
Home – We’re Finally Landing
Watch: • HOME – We’re Finally Landing [Synthwa…]
Stanley: An Overview
Introducing the DIY NVIDIA Jetson Nano Drone
Meet Stanley, your new gadget fascination and a testimony to what a little ingenuity and a lot of tech-savvy can accomplish. Stanley is not just any drone. Powered by the NVIDIA Jetson Nano and sporting an OAK-D Lite camera, he’s a pioneer in the exploration of in-flight AI research. This drone is designed for those who crave the thrill of DIY innovation, backed by cutting-edge technology. Let’s dive into what makes Stanley not just a flying machine, but a flying laboratory.
Purpose of In-Flight AI Research
The primary mission of Stanley is to serve as a conduit for in-flight AI experimentation. By marrying the NVIDIA Jetson Nano’s processing capabilities with the depth-sensing features of the OAK-D Lite camera, Stanley takes to the skies not just for show, but for serious AI research. The aim is to explore autonomous capabilities and AI integration, setting the stage for groundbreaking developments in how drones interact with their environments and perform complex tasks autonomously.
Key Components
Listing Essential Parts and Their Roles
Building Stanley required a curated selection of top-notch tech components. At the heart is the PixHawk 2.4.8, a reliable controller renowned among drone enthusiasts for its robustness. The NVIDIA Jetson Nano, prized for its AI processing prowess, complements the PixHawk. Stanley’s frame is the S550, a sturdy structure designed to withstand the rigors of flight. Other key components include the Ublox 7M GPS Module for navigation, HobbyPower 2212-KV920 Motors for propulsion, a LiPo Battery to power it all, SimonK ESCs, an F701 DSMX Receiver, and a Turnigy SBEC for power stabilization. Together, these parts form a symbiotic system aimed at achieving high performance and stability.
Overview of the NVIDIA Jetson Nano and OAK-D Lite Camera
The NVIDIA Jetson Nano is a compact yet potent AI computer, tailored for running multiple neural networks in parallel. It fits perfectly within the limited space of a drone, yet offers the kind of computational power normally reserved for larger machines. Paired with the OAK-D Lite camera, which excels at providing depth perception, this duo equips Stanley with the ability to process information about its surroundings in real-time. This configuration not only enhances Stanley’s ability to fly autonomously but also allows for sophisticated AI-driven decision-making during flight.
Building Stanley: The Process
Introduction to the Build Process
Creating Stanley wasn’t just a task—it was a journey paved with patience, curiosity, and a willingness to learn from past projects. This process started with a tear-down of old components and proceeded methodically through assembly, aligning the vision of a superior, more capable drone with the practicality of its execution.
Stages from Preparation to Assembly
The build was divided into clear stages: beginning with preparation, which included acquiring and organizing the essential components. Following preparation was soldering, an intricate phase where electrical connections were made robust and reliable. After ensuring the electronics were sound, frame assembly took precedence, slowly bringing Stanley to life physically. Lastly, all the components were meticulously assembled, ensuring everything fit harmoniously into Stanley’s compact frame.
Frame Assembly
Detailed Steps to Assemble the S550 Frame
Assembling the S550 frame was a task requiring precision and care. The process started with securing the motor arms to the power distribution board, vital for stability and power access. This was followed by installing the legs, ensuring they were firmly attached, not just to support the drone, but to cushion landings. The process was meticulous but rewarding, as piece by piece, Stanley began to take form.
Mounting Motor Arms and Crucial Connections
The motor arms required precise mounting to ensure stability in flight. Each arm was securely fastened, and wires routed meticulously to prevent entanglement. Connections were double-checked to ensure power would flow without interruption, setting the stage for electronics installation. This careful attention to detail during assembly would translate to better performance and less hassle during flight operations.
Electronics Installation
Integrating the PixHawk 2.4.8 and Jetson Nano
The integration of the PixHawk and Jetson Nano was a crucial phase. The PixHawk acted as Stanley’s brain, overseeing flight operations, while the Jetson Nano added a layer of intelligence, processing data, and executing AI models. This integration involved creating communication channels between these components, ensuring they worked in harmony to stabilize Stanley in flight while simultaneously running AI algorithms.
Ensuring Proper Connections and Stability
Achieving proper connections was paramount for avoiding mid-flight failures. This involved carefully soldering the ESCs to the power distribution board and ensuring all power paths were secure and reliable. Each cable and wire was routed for optimal connectivity, balancing stability with the need for flexibility in the drone’s compact interior. Such meticulous craftsmanship ensured that Stanley was not just functional but capable of handling the rigors of in-flight AI tasks.
Calibration Process
Setting Up the Transmitter and Emergency Stop
Calibration began with the configuration of the transmitter. This included setting up safety features, such as an emergency stop, crucial for preventing accidents. The transmitter was programmed to respond accurately to inputs, ensuring Stanley could be flown precisely and safely.
Testing and Adjusting Motor Spin Direction
Ensuring that each motor spun in the correct direction was critical for flight stability. Initial tests involved powering up the motors, observing their spin, and adjusting wiring as needed to reverse the direction of any incorrect spins. This step was vital, as improper motor installation could lead to airborne disasters, turning precious engineering hours into a destructive flight.
Challenges and Solutions
Encountering Mistakes and Learning from Them
Building Stanley wasn’t free of mistakes. Early mishaps, especially during motor installations, led to propellers flying off. These experiences, while initially frustrating, were invaluable learning opportunities. They highlighted the importance of thorough planning and double-checking every step. Each mistake prompted a reassessment and a subsequent improvement of assembly techniques.
Improving Soldering Skills and Motor Installations
Soldering, though seemingly simple, proved initially challenging. Missteps here included poor temperature control leading to weak joints. However, with practice and better equipment, these skills improved significantly. Similarly, motor installations initially led to frequent mistakes. Methodical re-approach and learning from detailed guides ensured more accurate installs in subsequent attempts.
Initial Test Flights
Testing Stanley in Adverse Conditions
Stanley had to prove his mettle in varied conditions. Test flights were conducted even in challenging weather, such as strong winds. These conditions tested all components’ durability and the build’s overall integrity, ensuring Stanley could handle less-than-ideal environments.
Addressing Flight Failures and Adjustments
Each test flight revealed areas needing adjustment. Initial attempts included instances where propeller detachment occurred mid-flight due to incorrect motor direction. Learning from these failures prompted careful calibration and re-testing, eventually resulting in stable and successful flights.
Future Aspirations for Stanley
Implementing Autonomous Follow Capabilities
Looking forward, Stanley aims to achieve autonomous following capabilities. By leveraging the computational power of the Jetson Nano in conjunction with the depth-sensing OAK-D Lite, future adaptations will include AI models capable of identifying and following targets autonomously. This capability heralds a new era of practical applications for drones in various domains.
Exploring AI Integration in Flight
Stanley is poised at the frontier of AI integration in flight. The journey doesn’t stop at autonomous following; future aspirations include more complex tasks such as real-time object detection and obstacle avoidance, all made possible through AI and machine learning models executed in-flight.
Conclusion
Reflecting on the Building and Testing Journey
The journey of Stanley’s creation was filled with learning and excitement. From initial design to overcoming assembly and calibration challenges, every step provided a lesson in patience and innovation. Witnessing Stanley take flight, handle computations, and advance AI research was a testament to the successful merging of creativity and technology.
Encouragement for DIY Drone Enthusiasts
For fellow DIY enthusiasts, Stanley’s journey underscores the potential of your own projects. Every stumble is a step toward mastery, and every misstep is a chance to grow. Dive into your projects with enthusiasm and curiosity, unafraid of mistakes, as they are the stepping stones to success. For those passionate about drones, this guide serves as both inspiration and a practical roadmap. Happy building, and may your skies be favorable!