Global ETD Search

1	Hanging Load Mock-Up For Experimental Setup To Test Flight System Stability Gundstedt, Anthony January 2023 (has links) Can we reduce disturbance, swing, and errors by delegating some of the tasks to a subsystem on the tool hanging from a drone to gain precision and dependability? This project covers developing and implementing a user-controlled prototype with an automatic control system and gripper for functionality and controlled movement in 5 DOF. AirForestry is developing a new way of working in the forestry business to provide a greener and healthier solution for the soil and the environment. The forestry industry emits approximately one million tons of CO2 per year from heavy diesel-fueled vehicles. With a renewable energy battery-powered drone of a diameter of 6.2 m and a hanging harvesting tool, they can significantly improve the environmental impact of forestry. The harvesting tool has an automatic control system and is user controlled to perform the thinning and cutting of trees. The main focus is to use IMU sensor data to produce accurate angle estimation to control the stability of a scaled prototype and to implement physical restrictions to make it able to be attached safely and tested on a drone but not limit the controlled mobility of the prototype with a system independent of the control system of the drone itself. The user can control the attitude and elevation and receive real-time sensor data wireless during operation. With the intended spacing of the attachment points on the drone, it can prevent unwanted swing from a push or displaced drop. It uses the accelerometer on the IMUsensor to calculate the angle of roll and pitch, although the accelerometer is sensitive to vibrations and rotations. The control system has a fast response and rise time, but it experiences noise and oscillations. Sensor fusion of the accelerometer and gyroscope in a complementary filter can be implemented to increase the accuracy of the angle estimations and decrease the noise, which will be reflected in the speed of the servos and thereby improve the stability and mobility of the system. drone hanging variable length mixer matrix PID controller microcontroller RC Telnet I2C FrSKY Sbus S.bus Forestry AirForestry Swing anti-swing IMU filter Control Engineering Reglerteknik
2	Through the Blur with Deep Learning : A Comparative Study Assessing Robustness in Visual Odometry Techniques Berglund, Alexander January 2023 (has links) In this thesis, the robustness of deep learning techniques in the field of visual odometry is investigated, with a specific focus on the impact of motion blur. A comparative study is conducted, evaluating the performance of state-of-the-art deep convolutional neural network methods, namely DF-VO and DytanVO, against ORB-SLAM3, a well-established non-deep-learning technique for visual simultaneous localization and mapping. The objective is to quantitatively assess the performance of these models as a function of motion blur. The evaluation is carried out on a custom synthetic dataset, which simulates a camera navigating through a forest environment. The dataset includes trajectories with varying degrees of motion blur, caused by camera translation, and optionally, pitch and yaw rotational noise. The results demonstrate that deep learning-based methods maintained robust performance despite the challenging conditions presented in the test data, while excessive blur lead to tracking failures in the geometric model. This suggests that the ability of deep neural network architectures to automatically learn hierarchical feature representations and capture complex, abstract features may enhance the robustness of deep learning-based visual odometry techniques in challenging conditions, compared to their geometric counterparts. artificial intelligence AI machine learning ML deep learning DL computer vision neural networks NN convolutional neural networks CNN visual odometry VO robustness motion blur AirForestry localization navigation ego-motion pose estimation SLAM DF-VO DytanVO ORB-SLAM3 artificiell intelligens maskininlärning datorseende Computer Sciences Datavetenskap (datalogi)

Search results

Hanging Load Mock-Up For Experimental Setup To Test Flight System Stability

Through the Blur with Deep Learning : A Comparative Study Assessing Robustness in Visual Odometry Techniques