The gondola attached to the parachute.
Stuck in a tree after successful flight.
Progression of the Paraglider

Introduction

Paragliding is an exciting sport where pilots can fly into the air and perform acrobatic tricks with only a parachute and a propeller. A full sized paraglider can be expensive and dangerous, so we are creating a small scale paraglider to bring the exciting sport to a more manageable size. Typical RC paragliders cost hundreds of dollars and offer little to no customization. By building the system from the ground up, it allows for a more cost effective approach with more flexibility on features that can be added.

The system architecture consists of both embedded and hardware aspects that work together to make the paraglider fly. At the heart of the embeded system is a BeagleBone Blue that controls a servo for steering and an Electric speed controller for power. The BeagleBone Blue has a built in IMU that is used for PID, and is powered by a Lipo battery. The hardware aspect consist of a parachute, DC Brushless motor and propellor, steering rods and a laser cut body to house everything.

Baseline Goals

The proposed milestones will form the foundation of the project.

Implement working electronics (Power Management, Motor Control and RC system)
Design and Build the Body
Have the Paraglider fly
Implement steering and stabilization

Reach Goals

The following goals are more complex and will be accomplished if time allows.

Implement powered flight
Increase range and duration of the flight
Decrease latency in the RC system
Create a stand-alone controller for the Paraglider

Alpha Prototype

For the first version of the paraglider the only electronic systems were the BeagleBone, Servo, and a 2s Lipo Battery. The BeagleBone communicated to a laptop over Wifi in order to control the steering of the paraglider from keyboard commands. On the hardware side the paraglider was made out of two pieces of acrylic to hold the electronics, and two wooden dowels that connected the acrylic body (also known as a Gondola) to the parachute.

After assembly was complete it was soon discovered that more was needed to make the paraglider fly. From a few test throws of the paraglider we discovered that the system needs to be propelled forward in order to sustain flight and it did not glide as we anticipated. Also the electronics were exposed and could be damaged in a high impact crash, but without a propeller we did not have to worry about this.

Bravo Prototype

For the second and third version of the paraglider we added a Brushless DC motor, ESC, and propeller along with a redesigned body in order to make another step towards flight. During this process we went through a few design iterations in the gondola design and replaced the wooden dowels with carbon fiber rods for a higher strength to weight ratio. We also replaced the smaller 2s Lipo battery with a beefed up 3s battery in order to increase the power output of our 1000KV motor. We also went from 6 inch to 8 inch propellers in order to generate more thrust for the system.

During our testing phase the paraglider crashed and destroyed itself on multiple occasions. Some crashes could be fixed with some minor fixes such as hot glueing pieces back together while a head on collision into a brick wall shattered all of the acrylic pieces on the paraglider (luckily the electronics survived). These tests prompted us to find a new testing location and change our material from 1/8 inch acrylic to 1/8 inch MDF.

Baseline Demo

For the Baseline Demo the Paraglider did achieve flight which was the primary goal of the project. After three iterations of the project the paraglider was in a mechanical state to fly, but still needed some tuning to control the flight and to increase the rang of the system. Our most successful flight at the time lasted for around five seconds before getting caught in a tree about 20 feet above the ground. Following the results of the baseline demo, we made the following goals to attempt for the reach demo. Baseline Demo Flight: https://youtu.be/cKtbJXNIHNs

Add a controller to eliminate the need for a laptop
Implement a Radio Controller instead of using wifi
Tune the mechanical design to stabilize flight by adjusting the strings
Program a "kill switch" into the system to protect the paraglider
Change the material of the design to make the gondola more robust.

Charlie Prototype

After the Baseline demo we continued to test the system and adjusted the strings of the paraglider in order for it to fly straight. The material of the gondola was upgraded to 1/8 inch ABS and prevented the paraglider from breaking upon landing. A controller was made using an ESP32 and 2 joysticks that communicated to the BeagleBone over UDP that allowed for control in the left and right directions along with altitude control. The controller was also programmed to turn off the paraglider when the pilot held the joystick in a certain position for a small amount of time which helped to prevent the propeller breaking upon landing.

Reach Demo

For the reach demo we were able to fly the paraglider for about 10 to 15 seconds with some control over the system. Mechanically the paraglider performed well flying hundreds of feet at a time. Our problem was that we could only communicate to the paraglider over wifi which only had around 30 feet of range at best. The paraglider would go out of range of the wifi and continue to fly until it timed out with each flight flying into a tree a few hundred feet away. At this stage our paraglider was more robust than the baseline version and could fly longer with some control, but still needed some tuning.

Final Product and Public Demo

After the reach demo we had a week to make some upgrades to the final product that drastically improved performance. The main problem that we had with our project was that wifi did not have enough range to control the paraglider which was resolved when we started using radio to communicate instead. Using a 915 MHz radio we were able to communicate a few hundred feet (up to around 1000 feet theoretically) and we never went out of range of the transceiver. The radio comunicated to the ESP32 over SPI, and we added an ESP32 to the paralider that communicated the the radio over SPI and then communicated the radio signals to the BeagleBone over UART. We could have eliminated the BeagleBone or ESP32 on the paraglider, but we only received the radios a few days before the demo and it was during finals week so time was limited.

For the public demo we were able to successfully fly the paraglider for about 20 seconds, and we could definitely go longer if we had more space. We are also limited by pilot expirence since even an off the self RC paraglider is not easy to fly. Overall we were satisfied with the result of this project where we developed the skills to make almost any other RC platform. Over the summer we plan to continue tuning the project by reducing components and possibly develop an open source RC platform that can be used for Drones, Cars, Boats, Paragliders, Hang-gliders, or any other small scale vehicle.

Software Effort

Paraglider software running on BeagleBone Blue and written in C using librobotcontrol library (For interfacing with BeagleBone IMU and PWM outputs)
Reads the yaw value from the IMU and uses that to adjust the servo position to try to stay flying in a straight line
Controller is based off an ESP32 and the code is written in C
Controller reads in joystick values with ADC and sends data to a 915 MHz radio
ESP32 reads in the data from the Radio over SPI and communicates to the Beaglebone over UART.

Technical Skills

Learned a about hardware design for flight (Keeping weight low, maximizing strength)
Mechanical manufacturing and materials
Choosing the proper brushless DC motor and ESC for an application
Communication over UDP
Communication over 915 MHz Radio
Programming the ESP32
Controlling Servos and ESCs from the BeagleBone
Multithreading on the BeagleBone