Examensarbeten för kandidatexamen
Länka till denna samling:
Browse
Senast publicerade
- PostEnhancing the MARV with Multi-Camera Support for Improved Remote Operation over 5G(2024) Bengtsson , Gabriel; Lithell, Anna; Ulstein, Oskar; Belal, Othman; Al-Wakkal, Shada; Chalmers tekniska högskola / Institutionen för elektroteknik; Chalmers University of Technology / Department of Electrical Engineering; Falkman, Petter; Lindström, ViktorAbstract The Marine Autonomous Research Vehicle (MARV) project aims to improve the efficiency of SSRS rescue missions. This report details the enhancement of the MARV with a multi-camera system for improved remote operation over a 5G network. The primary objective is to expand the field of view and enhance situational awareness through a system that supports a quad-camera setup in order to provide 360◦ coverage in future development of the MARV. The system streams high-definition video with low latency via the Secure Reliable Transport (SRT) protocol and GStreamer framework. A waterproof enclosure and connectivity infrastructure were installed, including a 5G router and network switch. The solution uses H.265 video compression for optimal quality under varying 5G network conditions. A user-friendly web-based interface was developed to be implemented in further development of the MARV.
- PostTidig karaktärisering av stroke genom videoanalys, maskininlärning och ögonspårning(2024) Ollila, Samuel; Ström, Eddie; Khatiri, Robin; Svensson, Teodor; Westerberg, Jacob; Chalmers tekniska högskola / Institutionen för elektroteknik; Chalmers University of Technology / Department of Electrical Engineering; Canderfjord, Stefan; Jalo, HoorStroke är en ledande orsak till dödlighet och funktionsnedsättning globalt. Snabb och tillförlitlig diagnos är avgörande för att optimera behandlingen, öka patientsäkerheten och rädda liv. Detta projekt syftar till att använda maskininlärningsalgoritmer för att bedöma misstänkta strokefall genom att tillämpa dem på data från ögonspårning genom videoanalys. Målet och det avsedda resultatet är att potentiellt förbättra effektiviteten i prehospital vård. National Institutes of Health Stroke Scale (NIHSS) är en skala som traditionellt har används för att klassificera stroke. Genom att digitalisera NIHSS och använda den som en mall för att identifiera ögonrörelseavvikelser, en vanlig indikator på stroke, hoppas man uppnå detta. I denna studie, på grund av bristen på patientdata, skapades en hybrid datamängd innehållande både verklig och syntetiska data. De verkliga videorna (n=99) bestod av friska individer som simulerade blickförlamning i enlighet med NIHSS-riktlinjerna. De syntetiska datan var nödvändig eftersom vissa ögonrörelser, som när en patient har problem med kranialnerv tre, är mycket svår att härma. Därför användes digitala animationsverktyg (Blender och After Effects) för att skapa videor med syntetiska ansiktet (n=65) som simulerar dessa symtom. Den riktiga datamängden validerades sedan av en strokespecialist. De algoritmer som valdes för att tolka de kombinerade datamängderna var convolutional neural networks (CNN), deep neural networks (DNN), gated recurrent units (GRU), support vector machines (SVM) and long short-term memory networks (LSTM). En hybrid datamängd utökade mängden träningsdata, en avgörande faktor för att förbättra tillförlitligheten hos alla maskininlärningsmodeller. LSTM uppnådde det bästa övergripande resultatet i studien och visade en noggrannhet på 88%, en känslighet på 87,7%, en specificitet på 94,1% och ett F1-värde på 86,7%, vilket understryker dess framtida potential som ett tillförlitligt diagnostiskt verktyg i prehospital miljö. Sammanfattningsvis visar resultaten att tillämpningen av maskininlärning och videoanalys för att digitalisera och klassificera strokeinducerade ögonrörelser erbjuder betydande fördelar. Denna teknik har potential att förändra och fungera som ett effektivt komplement till traditionella metoder för strokebedömning. Innan dessa tekniker kan implementeras i praktiken krävs dock ytterligare forskning och förfining av metoderna.
- PostDesigning a PCB microstrip antenna array for 5G mm-wave frequencies Designing and testing a dual-polarized aperture-coupled microstrip patch antenna for the (26.5-29.5) GHz band(2024) Eriksson, Simon; Kraamer, Johanna; Bujalla, Per Ingmar; Ali Shah, Syed Ahsan; Chalmers tekniska högskola / Institutionen för elektroteknik; Chalmers University of Technology / Department of Electrical Engineering; Ström, Erik; Vilenskiy, ArtemAbstract This report presents the development and testing of a dual linearly polarized patch antenna array. The designed antenna covers the frequency range of (26.5-29.5) GHz, which also is known as the 5G NR FR2 n257 band. The antenna went through multiple iterations of simulation in Ansys HFSS, starting with the construction of a single element that then served as a building block for a three element array. After completing the array in HFSS it was sent for manufacturing. The project compares the simulated results to the ones from the tested manufactured array. For the antenna to be effective in a communication scenario, targets were set for various performance aspects including the radiation pattern, total efficiency and scattering parameters (S-parameters). Each element in the array is aperture-coupled, cavity-backed and dual-polarized. The results showed that for a single isolated element, the peak broadside gain was 6.88 dBi for the vertical polarized port and 6.58 dBi for the horizontal polarized port. Moreover, the cross-polarization discrimination was > 14.6 dB, which suggest that it could be used for MIMO applications. It should be noted that these values are taken from the HFSS results as only the array and not the single element was manufactured. The manufactured antenna underwent multiple tests. The S-parameters were measured with a VNA, the radiation pattern was measured in an anechoic chamber and lastly the total efficiency was characterized in a reverberation chamber. All measured and simulated S-parameters achieved the target. Beam steering capabilities were also investigated by phase shifting each elements in post-processing from the accumulated data of the radiation pattern. The test results from the manufactured array follow the simulated ones well in its S-parameters and radiation pattern. The measured total efficiency was above > 75%. Finally, beam steering was achieved with a scanning range of ±44◦ for the simulated data and ±46◦ for the measured data.
- PostAnalyses of EMC- & Grid-connected Filters in an Electric Power-train(2024) Eriksson, Andreas; Ekborg, Tobias; Chalmers tekniska högskola / Institutionen för elektroteknik; Chalmers University of Technology / Department of Electrical Engineering; Maaskant, Rob; Andersson, EmilAbstract Epiroc have had a need to investigate the EMC and LCL-filters that are currently placed and operating in a lot of their vehicles. They have questioned whether or not their currently used capacitor and inductor-based filters are working adequately in filtering unwanted noise or if there could be a risk of wear to the system with prolonged operation and if any improvements could be made. This thesis has, with the help of Epiroc and extensive research, examined these questions in detail by conducting measurements on a drill rig vehicle at Epiroc’s facilities in Örebro. The thesis dives into areas such as Electromagnetic interference, electrical noise prevention methods and signal processing. The measurements indicated that both of the filters work and do filter unwanted noise, differences can be seen in the overall noise level. There are however some ideas outlined in this thesis that potentially could help increase the noise suppression of the filters.
- PostDORA - Dexterous Robot Assistant Autonomous door opening with a mobile manipulator(2024) Bramsved, Felix; Gunterberg-Klase, Lukas; Hendriksen, Vincent; Schuchert, Ellen; Chalmers tekniska högskola / Institutionen för elektroteknik; Chalmers University of Technology / Department of Electrical Engineering; Zou, ChangfuAbstract This thesis has focused on building a ROS based system for autonomous door detection and opening using a mobile manipulator. The mobile manipulator in question, DORA (Dexterous Robot Assistant), is based on the MiR200 mobile platform along with an UR10 robotic arm. The robot is equipped with an Intel realsense D435i depth camera. DORA had a navigation system and motion planning system for the arm already implemented and these systems were used in the process of this project. Object detection was achieved through a YOLO algorithm on the image stream from the depth camera, and pose estimation was done using the depth values obtained. The program was tested and found to be consistent and accurate when the robot was located in front of the door, but becomes unreliable in edge cases when the door is located at the edge of the field of vision or on an angle. The door opening program creates a trajectory for opening a door with unknown kinematics and an uncertain hinge location by utilizing readings from a force/torque sensor that is located between the arm and the end-effector, and the position of said end-effector. By employing the MoveIt framework, DORA was able to follow that trajectory and successfully open the door. Testing has confirmed that the force/torque sensor enables the program to operate successfully even in the presence of a 16% error in the estimation of the placement of the hinge position. In conclusion it has been determined that although further development in safety systems and increased versatility is required for real world use, a solid foundation for further work has still been created in this thesis and the capabilities of DORA to interact with the world has been greatly expanded.