A Review on Mixed Reality-Based Simulator for Drone Pilot Training

In recent years, drones have grown in popularity due to their capacity to carry out various activities faster and more successfully than conventional techniques. Drones are employed in a variety of industries, including photography, search and rescue, surveying, and agriculture. However, due to the sophisticated nature of the technology, operating a drone requires specialized knowledge and training. Training drone pilots can be difficult and expensive since it requires hands-on experience and regulatory and safety measures expertise. The use of simulators is one potential answer to the difficulties in training drone pilots. Simulators offer operators a safe, controlled environment where they may hone their abilities and acquire experience without having to worry about endangering their equipment or hurting others. The challenge at hand involves developing an immersive drone flight training system that seamlessly integrates with the real world, ensuring users experience a heightened sense of engagement without motion sickness. The simulation must employ physics that precisely mirror real drone flight dynamics, eschewing fictional models for authenticity. Central to this objective is the implementation of a user interface providing accurate telemetry values, affording pilots a comprehensive understanding of crucial realistic factors during flight. Additionally, the system demands a sophisticated training module that imparts practical skills and knowledge, ensuring proficiency in drone operation through real-world scenarios. The limitations of simulators, such as their resemblance to VR games rather than serious training environments, pose challenges for training drone pilots. Our research focuses on addressing these challenges by closely modelling the virtual environment to mirror the real environment using tools like Unity3D and Blender. Additionally, efforts have been made to model the physics of the drone by applying static and dynamic propeller thrust based on RPM data from the open-source flight planner ArduPilot. This work also aims to provide users with control by enabling them to switch between various views, including first person, third person, and fixed views. The testing procedure employed a hoop course system. This involved a skilled drone pilot navigating through progressively challenging courses, with factors such as reduced hoop size, intricate manoeuvres, increased hoop quantity, and diminished lighting contributing to heightened difficulty levels. The pilot's score and time served as the benchmark. Novice pilots then tested the application, with their scores and times recorded. As they approached the benchmark, course difficulty increased.