Quadrotor Platform Project

General Overview

MIL (Model-in-the-loop)

Plant

• 🔵 (planned) - Tunning of generic inner loop PID
• 🔵 (planned) - Make ground-truth available
• 🔵 (planned) - Finalize plant component
• 🔵 (planned) - Create linear models for futher systems identification and inner loop PID gains tuning
• ⚪ (Open) - Identify linear model based on real drone experiments
• ⚪ (Open) - Adjust inner loop PID based on real drones experiments
• ⚪ (Open) - Compare model identified with other real experiments
• ⚪ (Open) - Create operation states for possible control (taking-off, landed, landing, etc)
• ⚪ (Open) - Verify if model's interface matches with real hardware

Sensor

• ⚪ (Open) - Create IMU model
• ⚪ (Open) - Create barometer model
• ⚪ (Open) - Create camera model

Localization

• ⚪ (Open) - Simulate tag localization package
• ⚪ (Open) - Simulate opt-track localization package
• ⚪ (Open) - Simulate ORB-SLAM package
• ⚪ (Open) - Investigate machine learning techniques for localization

Control

• ⚪ (Open) - Implement PID position control
• ⚪ (Open) - Implement LQR position control
• ⚪ (Open) - Implement Model Preditive Control
• ⚪ (Open) - Investigate machine learning technique for control

Path planning

• ⚪ (Open) - React to gesture commands
• ⚪ (Open) - React to auditive commands
• ⚪ (Open) - investigate object detection techniques
• ⚪ (Open) - Investigate object search for techniques

SIL (Software-in-the-loop)

Plant

• ⚪ (Open) - Create robot description
• ⚪ (Open) - Create gazebo plugin based on MIL plant model
• ⚪ (Open) - Compare experiments with MIL and real experiments

Sensor

• ⚪ (Open) - Add IMU sensor gazebo plugins to robot
• ⚪ (Open) - Add barometer sensor gazebo plugin to robot
• ⚪ (Open) - Add camera sensor gazebo plugin to robot

HIL (Hardware-in-the-loop)

Plant

• ⚪ (Open) - Summarize available interface with Tello Drone
• ⚪ (Open) - Create basic interface to send and receive commands from/to the drone
• ⚪ (Open) - Define experiments to model identification
• ⚪ (Open) - Apply model identification experiments and save data

Summary

MIL (Model-in-the-loop)

Plant

Tunning of generic inner loop PID

Platform: MIL | Component: Plant | Time spent: 1m

teste

Make ground-truth available

Platform: MIL | Component: Plant | Time spent:

Disponibilizar ground-truth

Finalize plant component

Platform: MIL | Component: Plant | Time spent:

finalizar bloco da planta

Create linear models for futher systems identification and inner loop PID gains tuning

Platform: MIL | Component: Plant | Time spent:

Criar os modelos lineares para identificação e ajuste dos ganhos PID

Identify linear model based on real drone experiments

Platform: MIL | Component: Plant | Time spent:

Levantar modelo linear com base em experimentos com drone real

Adjust inner loop PID based on real drones experiments

Platform: MIL | Component: Plant | Time spent:

Ajustar controladores com base no hardware

Compare model identified with other real experiments

Platform: MIL | Component: Plant | Time spent:

Comparar o modelo corrigido com experimentos reais

Create operation states for possible control (taking-off, landed, landing, etc)

Platform: MIL | Component: Plant | Time spent:

Criar estados de operação para possível controle (voando, pousado, taking off, etc)

Verify if model's interface matches with real hardware

Platform: MIL | Component: Plant | Time spent:

Verify if model's interface matches with real hardware

Sensor

Create IMU model

Platform: MIL | Component: Sensor | Time spent:

Criar modelo de IMU

Create barometer model

Platform: MIL | Component: Sensor | Time spent:

Criar modelo de barômetro

Create camera model

Platform: MIL | Component: Sensor | Time spent:

Create camera model

Localization

Simulate tag localization package

Platform: MIL | Component: Localization | Time spent:

Localização por tag

Simulate opt-track localization package

Platform: MIL | Component: Localization | Time spent:

Opt-track

Simulate ORB-SLAM package

Platform: MIL | Component: Localization | Time spent:

ORB-SLAM

Investigate machine learning techniques for localization

Platform: MIL | Component: Localization | Time spent:

Investigar técnicas avançadas de ML para localização

Control

Implement PID position control

Platform: MIL | Component: Control | Time spent:

PID

Implement LQR position control

Platform: MIL | Component: Control | Time spent:

LQR

Implement Model Preditive Control

Platform: MIL | Component: Control | Time spent:

MPC

Investigate machine learning technique for control

Platform: MIL | Component: Control | Time spent:

Tecnica de Machine Learning?

Path planning

React to gesture commands

Platform: MIL | Component: Path planning | Time spent:

Reagir a comandos visuais ou gestuais

React to auditive commands

Platform: MIL | Component: Path planning | Time spent:

Reagir a comandos auditivos

investigate object detection techniques

Platform: MIL | Component: Path planning | Time spent:

Implementar técnica de identificação de objetos

Investigate object search for techniques

Platform: MIL | Component: Path planning | Time spent:

Implementar técnica de procurar por objeto

SIL (Software-in-the-loop)

Plant

Create robot description

Platform: SIL | Component: Plant | Time spent:

Create robot description

Create gazebo plugin based on MIL plant model

Platform: SIL | Component: Plant | Time spent:

Create gazebo plugin based on MIL plant model

Compare experiments with MIL and real experiments

Platform: SIL | Component: Plant | Time spent:

Compare experiments with MIL and real experiments

Sensor

Add IMU sensor gazebo plugins to robot

Platform: SIL | Component: Sensor | Time spent:

Add sensor gazebo plugins to robot

Add barometer sensor gazebo plugin to robot

Platform: SIL | Component: Sensor | Time spent:

Add barometer sensor gazebo plugin to robot

Add camera sensor gazebo plugin to robot

Platform: SIL | Component: Sensor | Time spent:

Add camera sensor gazebo plugin to robot

HIL (Hardware-in-the-loop)

Plant

Summarize available interface with Tello Drone

Platform: HIL | Component: Plant | Time spent:

Summarize available interface with Tello Drone

Create basic interface to send and receive commands from/to the drone

Platform: HIL | Component: Plant | Time spent:

Create basic interface to send and receive commands from/to the drone

Define experiments to model identification

Platform: HIL | Component: Plant | Time spent:

Define experiments to model identification

Apply model identification experiments and save data

Platform: HIL | Component: Plant | Time spent:

Apply model identification experiments and save data