Automated Helicopter

PROJECT CONCEPT

This projects goal is to create a system to allow for automated take-off and landing of a model helicoptor that is separate from the actual helicoptor. This would allow for the control system to be modular and applicable to other flying device. By using accelerometers and a microcontroller we hope to convert a Remote Control (RC) helicopter to automatically complete certain tasks that would normally require a human operator.

MOTIVATION

Motivating for this effort is that autonomous vehicle research has been going on for a long time, but consumer products are few an far between; cost being a large prohibitive factor. Furthermore, most research in this field has been focus mainly on military applications; consumer use is largely unexplored.

COMPETITIVE ANALYSIS

Draganflyer X6; http://www.draganfly.com/uav-helicopter/draganflyer-x6/

This three rotor helicoptor is an incredible machine that uses 11 sensors to self-stabilize and provide the user an ability to set the device to operate semi-autonomously. Cost is a gigantic 15,000 USD.

TECHNICAL SPECIFICATIONS

Helicopter

:: Ideal

- Genesis 450PE RTF $329.00 || http://www.xheli.com/fulo45rtfhec.html
- Anti-Crash Kit $6

Sensor suite package

Sensors and Controller

:: UAV v2 Development Platform - $299.95 || sku: GPS-09038 || dsPIC30F4011 chip, MMA7260 three axis accelerometer,  3 ADXRS401 gyros

REQUIREMENTS

-Develop a stability control system that allows the helicopter to maintain a hover without input from the pilot.

Steps taken:

1) Wire the uC to manipulate the helicopter servos based on the uC’s code; connect the control of the helicopter to the processor.

2) Monitor the sensor data input into the uC and demonstrate the ability to react to this input.

3) Process the data accordingly translate it into the correct PWM signal.

4) Test helicopter’s flight ability.

ARCHITECTURE

Diagram of Stability Control System

USE CASES (INTERACTION DIAGRAMS)

To be completed.

SYSTEM STATES & TRANSITIONS

State Machine

RISKS & MITIGATION STRATEGIES

1. The mechanical system noise from the helicopter body is too high and noise from the sensors themselves.
   -attenuate the system noise by utilizing a butterworth filter and sanitizing the sensors’ inputs.
   -reduce the mechanical noise by fixing the mechanical parts, such as straightening out the rotor shaft.
2. Mechanical failures of the helicopter.
   -examine the mechanical parts after every test flight. Make sure no loosened screws .
3. Possible helicopter crash during the flight.
   -attach anti crash kit to the helicopter.

ERROR HANDLING

Errors to be monitored and corrected:

1. Position of helicopter, and how it deviates from the position needed to maintain a hover.

-Use PD controller to control helicopter swashplate

2. Difference in orientation as sensed by the gyropscopes and accelerometers; gyroscopes are precise, accelerometers are accurate.

-Use PI controller to correct for the difference in data before being sent to the PD controller (point 1).

3. Sensors errors: incorrectly calibrated sensor board, or mechanical noise that affects the sensors

-Mechanically correct placement with spring dampeners, software corrected calibration, and aggressive data filtering.

IMPLEMENTATION DETAILS

*See Architecture section for visual details.
As the goal of the project was to create a stability control system, the board needed to be mounted to the helicopter to give a good sense of the helicopter’s position. Springs were used to help dampen high frequency mechanical noise.

Many PWM inputs and outputs are used
The board is mounted upside down

As there are only 3 output PWM lines, we only configured the board to control roll and pitch with the helicopter’s main rotor swashplate which is controlled by 3 servos. The rotor speed control, main rotor blade pitch (which controls lift), and tail rotor blade pitch (which controls yaw) were left under manual control with the remote transmitter. With some configurations of the code, full manual control can be assumed with the remote.

USB, anyone?

Because data needed to be collected from the board in real time, a serial-to-USB converter was used along with a usart from the sensor board. This allowed us to gather sensor data for use in configuring our filters.

TEST CASES

1. Have the helicopter hover under total manual control. This is how the mechanics of the helicopter can be tested: if the helictoper is mechanically too unstable to reach a hover with manual control, it is probably not safe to fly with automatic control.
2. Ability to maintain a stable hover without significant roll and pitch, and hopefully without significant lateral movement in a sanitized environment–this means without external affectors such as wind.
3. Realistic real-world testing: allow for wind or simulate a change in stead state by perturbing the system.

EXPERIMENTAL EVALUATION

GRAPHS OF FILTERED DATA
VIDEO OF FLIGHT…. SUCCESSFUL FLIGHT PLZ

LESSONS LEARNED

Helicopters like to crash. They have a penchant for it. Crashing is bad. Bad helicopter, bad.
CAUTION: the following image is quite graphic and might disturb some helicopters that can still fly.

Dead Helicopter Flying

We have also learned that Zip-ties are some of the best engineered slices of perfection the world over. Duct tape and hot glue is for amatures.

FUN STUFF

Fun stuff, indeed.

REFERENCES

Roses are red,
Violets are blue.
Look at our poster,
in ppt, pdf, and png, too!

A movie? Yes.