Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Creation of a Scalable Hands-on Quadrotor Activity for the Control Systems 1 Class
The goal of this project is to remove the need for a motion capture system from the hands-on quadrotor control P&S offered by the IfA and design a classroom activity that provides a reliable learning experience for implementing one or more of the concepts taught in the Control Systems 1 course.
The advantage of using a motion capture system is that it provides very precise measurements, thus allowing a reliable learning experience for the students taking the P&S, even for complex manoeuvres and interactions. The disadvantage of a motion capture system is that it is not readily portable or scalable due to cost and space constraints. To achieve the goal of this project, the position and orientation will be obtained through measurements from sensors available on-board a Crazyflie 2.1 quadrotor. This will allow a short version of the quadrotor P&S to be conducted in virtually any location, without the need for additional infrastructure, and with the class size limited only by the number of quadrotors available. The ultimate goal for class size is to have the entire Control System 1 class performing quadrotor experiments simultaneously.
The Crazyflie 2.1 is ideal for this project because it comes with a range sensor for vertical height, an optical flow sensor for changes in (x,y) position, and an inertial measurement unit (IMU) for estimation of the orientation. However, these sensors are strongly coupled and less accurate than a motion capture system. The firmware of the Crazyflie 2.1 is open source and already implements position and orientation estimation from the sensors described. Based on initial testing, it is expected that further filtering, outlier detection, and safety mechanism design is needed to make the system sufficiently reliable for usage in a classroom setting.
As the measurements are taken onboard the Crazyflie, it is required to develop a software archi- tecture that differs significantly from that currently used for the P&S. Two high level architectures are envisioned: 1) where the sensor measurements are sent from the Crazyflie to the student’s laptop, the control algorithm is executed on the laptop, and the computed commands are sent back to the Crazyflie, or 2) where all control computations are performed onboard the Crazyflie and the software is setup so that a student can quickly re-programm the control algorithm and is able to change parameters mid-flight.
For the classroom activity, we envision a 2-3 hour hands-on session, with some additional pre-work, offered as an optional activity to the Control System 1 participants in future years. Beyond this, the students undertaking this project will be given the freedom to propose and design a classroom activity that they consider most motivating and informative. The reason is that as a student you provide a unique, relevant, and fresh perspective on what hands-on activity assists your fellow students to achieve a deeper understanding and insight of the control technique implemented.
The advantage of using a motion capture system is that it provides very precise measurements, thus allowing a reliable learning experience for the students taking the P&S, even for complex manoeuvres and interactions. The disadvantage of a motion capture system is that it is not readily portable or scalable due to cost and space constraints. To achieve the goal of this project, the position and orientation will be obtained through measurements from sensors available on-board a Crazyflie 2.1 quadrotor. This will allow a short version of the quadrotor P&S to be conducted in virtually any location, without the need for additional infrastructure, and with the class size limited only by the number of quadrotors available. The ultimate goal for class size is to have the entire Control System 1 class performing quadrotor experiments simultaneously.
The Crazyflie 2.1 is ideal for this project because it comes with a range sensor for vertical height, an optical flow sensor for changes in (x,y) position, and an inertial measurement unit (IMU) for estimation of the orientation. However, these sensors are strongly coupled and less accurate than a motion capture system. The firmware of the Crazyflie 2.1 is open source and already implements position and orientation estimation from the sensors described. Based on initial testing, it is expected that further filtering, outlier detection, and safety mechanism design is needed to make the system sufficiently reliable for usage in a classroom setting.
As the measurements are taken onboard the Crazyflie, it is required to develop a software archi- tecture that differs significantly from that currently used for the P&S. Two high level architectures are envisioned: 1) where the sensor measurements are sent from the Crazyflie to the student’s laptop, the control algorithm is executed on the laptop, and the computed commands are sent back to the Crazyflie, or 2) where all control computations are performed onboard the Crazyflie and the software is setup so that a student can quickly re-programm the control algorithm and is able to change parameters mid-flight.
For the classroom activity, we envision a 2-3 hour hands-on session, with some additional pre-work, offered as an optional activity to the Control System 1 participants in future years. Beyond this, the students undertaking this project will be given the freedom to propose and design a classroom activity that they consider most motivating and informative. The reason is that as a student you provide a unique, relevant, and fresh perspective on what hands-on activity assists your fellow students to achieve a deeper understanding and insight of the control technique implemented.
The goal of this project is to remove the need for a motion capture system from the hands-on quadrotor control P&S offered by the IfA and design a classroom activity that provides a reliable learning experience for implementing one or more of the concepts taught in the Control Systems 1 course. Achieving this goal requires the following tasks to be completed:
1. Familiarise with the existing quadrotor P& Scourse content, and the software used for the hands-on component of the course.
2. Develop a patch for the Crazyflie 2.1 firmware that integrates it with the existing offboard software used for the P&S.
3. Log and analyze the sensor measurement available from the Crazyflie. This will likely require the development and implementation of estimation algorithms, outlier detection, and comparison with Vicon for verification purposes.
4. Develop and test, the most promising of the two architectures envisioned (or propose a new architecture based on the learning from tasks 1-3).
5. Propose and develop a classroom activity that takes concepts from the Control Systems 1 course and provides a hands-on implementation of this concept on the system developed in tasks 1-4.
6. Develop the necessary interfaces, documentation, and safety mechanisms so that the architecture from task 4 and the classroom from task 5 are sufficiently reliable, user-friendly, and scalable.
The goal of this project is to remove the need for a motion capture system from the hands-on quadrotor control P&S offered by the IfA and design a classroom activity that provides a reliable learning experience for implementing one or more of the concepts taught in the Control Systems 1 course. Achieving this goal requires the following tasks to be completed:
1. Familiarise with the existing quadrotor P& Scourse content, and the software used for the hands-on component of the course.
2. Develop a patch for the Crazyflie 2.1 firmware that integrates it with the existing offboard software used for the P&S.
3. Log and analyze the sensor measurement available from the Crazyflie. This will likely require the development and implementation of estimation algorithms, outlier detection, and comparison with Vicon for verification purposes.
4. Develop and test, the most promising of the two architectures envisioned (or propose a new architecture based on the learning from tasks 1-3).
5. Propose and develop a classroom activity that takes concepts from the Control Systems 1 course and provides a hands-on implementation of this concept on the system developed in tasks 1-4.
6. Develop the necessary interfaces, documentation, and safety mechanisms so that the architecture from task 4 and the classroom from task 5 are sufficiently reliable, user-friendly, and scalable.