The combination of machine learning on embedded systems has quickly increased throughout the years. Subsets like TinyML have become an integral part of how embedded systems implement machine learning. The field has evolved quickly, and TinyOL is an emerging subset that redefines what is possible with embedded systems. This report presents a comparison of how a neural network that implements incremental online learning learns and adapts how to do simple tasks in home automation. The comparison is done with another system, mainly a proportional-integral-derivative (PID). The systems were tasked with controlling an LED lightning threshold based on feedback from the user. The systems were evaluated based on their mean absolute error (MAE) and accuracy in predicting the output of the LED lighting system. The MAE values of both systems were compared for different target outputs and threshold values, and the accuracy was calculated by comparing the number of successful iterations to the total number of iterations. The results show that the neural network has an accuracy of 50\% when a learning rate of 0.2 is used, 97.5\% when a learning rate of 0.5 is used, and 47.5\% when a learning rate of 1.0 is used. The PID control system had accuracy values of 45\% when using an adaption rate of 0.2, 47.5\% when using an adaption rate of 0.5, and 90\% when using an adaption rate of 1.0. The neural network also showcased a lower median MAE for every test conducted. The study provides insights into the effectiveness of different control systems and can inform the development of similar systems in the future. / Kombinationen av maskininlärning på inbygga system har snabbt ökat under åren. Tekniker som TinyML har snabbt blivit en integrerad del av hur inbyggda system implementerar maskininlärning. Teknikerna har snabbt utvecklats och TinyOL är en framväxande delmängd av TinyML som omdefinierar vad som är möjligt med inbyggda system. Denna rapport presenterar en jämförelse av hur ett neuralt nätverk som implementerar inkrementell online-inlärning lär och anpassar sig för att utföra enklare uppgifter inom hemautomation. Jämförelsen görs med ett annat system, huvudsakligen en proportional-integral-derivative (PID). Systemen fick i uppgift att kontrollera en LED-lampa baserat på användarens feedback. Systemet utvärderas baserat på deras mean absolute error (MAE) och noggrannhet i att förutsäga börvärdet för LED-belysningen. MAE-värderna för båda systemen jämfördes för de olika målen och börvärdena, och noggrannheten beräknades genom att jämföra antalet lyckade iterationer med det totala antalet iterationer. Resultaten visar att neurala nätverket har en  noggrannhet på  50\% när en learning rate på 0.2 användes, 97.5\% när en learning rate på 0.5 användes och 47.5\% när en learning rate på 1.0 användes. PID kontroll system hade en noggranhet på 45\% när en adaption rate på 0.2 användes, 47.5\% när en adaption rate på 0.5 användes och 90\% när en adaption rate på 1.0 användes. Det neurala nätverket visade också ett lägre MAE-värde på alla de testerna som utfördes. Studien ger insikter i effektiviteten hos olika kontrollsystem och kan hjälpa utvecklingen av liknande system i framtiden.

The development of advanced embedded systems requires a systematic approach as well as advanced tool support in dealing with their increasing complexity. This complexity is due to the increasing functionality that is implemented in embedded systems and stringent (and conflicting) requirements placed upon such systems from various stakeholders. The corresponding system development involves several specialists employing different modeling languages and tools. Integrating their work and the results thereof then becomes a challenge. In order to facilitate system architecting and design integration of different models, an approach that provides dedicated workspaces/views supported by structured information management and information exchange between domain models and tools is required. This work is delimited to the context of embedded systems design and taking a model based approach. The goal of the work is to study possible technical solutions for integrating different models and tools, and to develop knowledge, support methods and a prototype tool platform. To this end, this thesis examines a number of approaches that focus on the integration of multiple models and tools. Selected approaches are compared and characterized, and the basic mechanisms for integration are identified. Several scenarios are identified and further investigated in case studies. Two case studies have been performed with model transformations as focus. In the first one, integration of Matlab/Simulink® and UML2 are discussed with respect to the motivations, technical possibilities, and challenges. A preliminary mapping strategy, connecting a subset of concepts and constructs of Matlab/Simulink® and UML2, is presented together with a prototype implementation in the Eclipse environment. The second case study aims to enable safety analysis based on system design models in a UML description. A safety analysis tool, HiP-HOPS (Hierarchically Performed Hazard Origin and Propagation Studies), is partially integrated with a UML tool where an EAST-ADL2 based architecture model is developed. The experience and lessons learned from the experiments are reported in this thesis. Multiple specific views are involved in the development of embedded systems. This thesis has studied the integration between system architecture design, function development and safety analysis through using UML tools, Matlab/Simulink, and HiP-HOPS. The results indicate that model transformations provide a feasible and promising solution for integrating multiple models and tools. The contributions are believed to be valid for a large class of advanced embedded systems. However, the developed transformations so far are not really scalable. A systematic approach for efficient development of model transformations is desired to standardize the design process and reuse developed transformations. To this end, future studies will be carried out to develop guidelines for model and tool integration and to provide support for structured information at both meta level and instance level. / QC 20101116

Introduction: This paper presented an analysis of the energy consumption and responsiveness of MicroPython in an embedded system. The purpose of this study was to understand the energy consumption and response time of a MicroPython based system to optimize its overall performance and efficiency. Two research questions had been formulated to concretize the purpose of this thesis: [RQ1] How does the energy consumption of a MicroPython based embedded system compare to that of a C-based embedded system for tasks utilizing low power modes? [RQ2] What is the wake-up response time of MicroPython for low power modes when receiving external and internal interrupts, and how does it compare to an established language like C on an embedded system? Method: To answer the research questions and achieve the purpose, an experimental study was conducted. The energy consumption of the MicroPython based system was analyzed under different scenarios. The time it took for MicroPython to respond to an interrupt request from a sleeping state was also measured. The data collected from the experiment was analyzed to determine the level of energy consumption and responsiveness of MicroPython in an embedded system. Results: The results indicated that C was generally more energy efficient and responsive than MicroPython for tasks utilizing low power modes for the Deepsleep mode. Although MicroPython proved to have shorter response times for the Lightsleep low power mode. For energy consumption, C was more stable in the measurements while MicroPython reached both lower minimum currents and higher maximum currents. Conclusions: In conclusion, this study found that while MicroPython could achieve lower power levels than C in both low power modes tested, it reached higher current levels upon waking up. Despite this, MicroPython could still be a choice for applications that spend longer durations in low power modes, as this could offset the increased current spikes during wake-up. Response times for MicroPython were faster than C in the Lightsleep internal interrupt case, but MicroPython exhibited significantly longer response times in the Deepsleep mode due to the system resetting and restarting the interpreter. Keywords: Embedded systems, Energy consumption, Interrupt requests, Low power modes, MicroPython, Responsiveness.

Unmanned aerial vehicles (UAVs), commonly known as drones, have revolutionized numerous fields ranging from aerial photography to surveillance and logistics. Achieving stable flight is essential for their successful operation, ensuring accurate data acquisition, reliable manoeuvring, and safe operation. This thesis explores the feasibility of employing a frontal mono camera and sensor fusion techniques to enhance drone stability during flight. The objective of this research is to investigate whether a frontal mono camera, combined with sensor fusion algorithms, can be used to effectively stabilize a drone in various flight scenarios. By leveraging machine vision techniques and integrating data from onboard gyroscopes, the proposed approach aims to provide real-time feedback for controlling the drone. The methodology for this study involves the Crazyflie 2.1 drone platform equipped with a frontal camera and an Inertial Measurement Unit (IMU). The drone’s flight data, including position, orientation, and velocity, is continuously monitored and analyzed using Kalman Filter (KF). This algorithm processes the data from the camera and the IMU to estimate the drone’s state accurately. Based on these estimates, corrective commands are generated and sent to the drone’s control system to maintain stability. To evaluate the effectiveness of the proposed system, a series of flight tests are conducted under different environmental conditions and flight manoeuvres. Performance metrics such as drift, level of oscillations, and overall flight stability are analyzed and compared against baseline experiments with conventional stabilization methods. Additional simulated tests are carried out to study the effect of the communication delay. The expected outcomes of this research will contribute to the advancement of drone stability systems. If successful, the implementation of a frontal camera and sensor fusion can provide a cost-effective and lightweight solution for stabilizing drones.

No description available.

In the current studied literature there is a lack of data and documentation on how outside factors affect the performance of ANC headphones. The purpose of this thesis is to investigate how these factors affect ANC headphone performance. The investigation into these factors was done by gathering quantitative data with experiments. The analysis method chosen was an ANOVA-inspired method that compared the differences found in the processed data. The results show that the sample variation between the number of speakers used did not differ by much. However, the variability in the samples collected using different noise types was noticeably higher while short noise resulted in odd-shaped graphs. When combining parameters from the different tests, effects such as graph shapes, seem to remain. These are the conclusions made. When testing with pink noise, the amount of speakers used and how long the pink noise is played makes a minimal difference for attenuation and variation. However, having more speakers and playing the sound for longer durations creates slightly more stable measurements. Unstable noise recordings are unsuitable for product quality testing due to the unreliable results they give. These noise recordings are still suitable for other research purposes.

The vehicle industry must reduce emissions to lessen their environmental impact. Using electric motors for propulsion and batteries for energy storage decreases the pollution and greenhouse gases emitted by heavy commercial vehicles such as trucks and buses. As they often travel long distances, a large energy storage is needed, and it needs to be possible to recharge it quickly. The fast recharging requires high power chargers with high voltage and high current, and that in turn necessitates a suitable overcurrent protection system to avoid damageif a fault would occur. Possible technologies for this overcurrent protection system are investigated in this thesis, and the different solutions are evaluated based on important metrics. A solid state breaker is found to be a promising technology that limits the current quickly and reliably. The evaluation including computer simulations shows that the technology is feasible, but some aspects of the circuit design need to be executed carefully like the inductance, the sampling frequency of the current measurement, and the design of the snubber circuit.

Particle suspension is a phenomenon when solid particles get trapped in a liquid. This phenomenon is a common occurrence in the water treatment industry. It is typically measured continuously throughout the treatment process to ultimately ensure high water-quality. Cerlic Controls AB specializes in developing sensors for the water treatment industry and have been doing so since 1977. Cerlic Control's sensors are used within the various stages of water-purification. The type of sensor central to this project uses transmission of light to determine a concentration of suspended particles. This project describes the development-process for one of Cerlic Controls existing embedded systems. The embedded system is undergoing a modernization process where the electronics and software are updated to a more modern design. This report describes the improvements to the measurement range while maintaining the same or improved resolution. These improvements are mostly performed by modern electronics and more advanced signal processing techniques in comparison to the previous implementation. A large portion of the previous implementation's software was split over two embedded systems due to limitations in processing power. This modern version focuses on bringing these two systems into one. The sensor's microprocessor has been replaced with a more modern higher performance alternative which allows all computations to be performed in the sensor. Linearity and stability are key aspects defining system performance. Optical filters block a certain amount of light and are heavily used in verification of performance. Suspensions using activated carbon were also used to evaluate the measurement range. The results of the project were largely successful. An initial implementation was achieved with optimistic results, but further work must be done before the sensor is ready for use in industry.

A current meter has been designed, intended for surveillance of low-power battery driven devices with a large dynamic range in their current consumption. Among such devices may, for example, be nodes in a wireless sensor network (WSN). The large dynamic range originates in the nodes’ states from an active mode for data collection and communication to a sleep mode for preserving very limited energy and prolonging the lifetime of the device. The active mode draws current up to hundreds of milliamperes whilst the sleep mode’s current may be as low as 22 nA. This makes the dynamic range larger than 6 orders of magnitude. To perform measurements that prevail over noise sources and resolution of involved components, an increase in the SNR must be made. If the current is measured with a shunt resistor, this may be done by varying its resistance as required. The resistance should be as high as possible to maximize the SNR, but low enough such that the burden voltage of the current meter is kept within tolerances of the device. A literature study was done, and with the gained knowledge a circuit was designed and simulated in LTSpice. Then a printed circuit board (PCB) of the circuit was designed and implemented using Altium Designer (an ECAD program), manufactured and hand-soldered. The final PCB was tested but shown to have some design flaws that deteriorated the performance, for example too frequent switching of the current range.

The rapid progress made in integrating enormous numbers of transistors on a single chip is making it possible for hardware designers to implement more complex hardware architectures in their designs. Nowadays digital telecommunication systems are implementing several forms of SOC (System-On-Chip) structures. These SOCs usually contain a microprocessor, several DSP cores (Digital-Signal-Processors), other hardware blocks, on-chip memories and peripherals. As new IC process technologies are deployed, with decreasing geometrical dimensions, the probabilities of hardware faults to occur during operation are increasing. Testing SOCs is becoming a very complex issue due to the increasing complexity of the design and the increasing need of a test mechanism that is able to achieve acceptable fault coverage in a short test application time with low power consumption without the use of external logic testers. As a part of the overall test strategy for a SOC, functional self-testing of a DSP core is considered in this project to be applied in the field. This test is used to verify whether fault indications in systems are caused by permanent hardware faults in the DSP. If so, the DSP where the fault is located needs to be taken out of operation, and the board it sits on will be later replaced. If not, the operational state can be restored, and the system will become fully functional again. The main purpose of this project is to develop a functional self-test of a DSP core, and to evaluate the characteristics of the test. This project also involves proposing a scheme on how to apply a functional test on a DSP core in an embedded environment, and how to retrieve results from the test. The test program shall run at system speed. To develop and measure the quality of the test program, two different coverage metrics were used. The first is the code coverage metric achieved by simulating the test program on the RTL representation of the DSP. The second metric used was the fault coverage achieved. The fault coverage of the test was calculated using a commercial Fault Simulator working on a gate-level representation of the DSP. The results achieved in this report show that this proposed approach can achieve acceptable levels of fault coverage in short execution time without the need for external testers which makes it possible to perform the self-test in the field. This approach has the unique property of not requiring any hardware modifications in the DSP design, and the ability of testing several DSPs in parallel.