Robotic Arm Trajectory Based on Machine Learning for tracking the wound

Résumé

Abstract Due to uncontrolled bleeding in victims and frequently delayed emergency response times to road accidents had frequently causes severe casualties. To solve this problem, the robotic arm used in this paper, based on machine learning, is programmed to monitor and react to bleeding in accident victims. A clotting agent is applied to the wounds by the robotic arm at the touch of a button, stopping the bleeding and possibly saving lives. The system's image processing and wound classification were done in a Python-based Jupyter environment. SolidWorks was used to design the robotic arm, which was programmed to respond to input from a Python model. The robotic arm's kinematics was tested in the Simulink environment, displaying the joints' successful operation. The robotic arm design was further analyzed in the Simcape environment, where the arm's trajectory was evaluated. The results showed that the robotic arm could effectively move toward the detected wound, following the training trajectory. Object tracking was successfully performed in Matlab under the Simcape environment, as evidenced by the alignment between the expected and actual trajectory line graphs. The study concludes that the machine learning-based robotic arm can accurately move to any desired position within its workspace, enabling the precise application of the clotting agent on the wound. This system holds significant potential for improving emergency response times and outcomes in road accidents.

Citer ce document

Accès au document

Texte intégral en lecture en ligne, réservé aux abonnés SPHAERO et aux membres de l'institution. Se connecter

Auteur(s)

Statistiques

Consultations : 1

Téléchargements : 0