Chalmers Open Digital Repository

Välkommen till Chalmers öppna digitala arkiv!

Här hittar du:

Studentarbeten utgivna på lärosätet, såväl kandidatarbeten som examensarbeten på grund- och masternivå
Digitala specialsamlingar, som t ex Chalmers modellkammare
Utvalda projektrapporter

Forskningspublikationer, rapporter och avhandlingar hittar du i research.chalmers.se

Enheter i Chalmers ODR

Välj en enhet för att se alla samlingar.

Visar 1 - 2 av 2

Senast publicerade

Circular food strategies for the Gothenburg circular economy action plan
(2026) NATH, RAVEENA
Resource efficiency can reduce material demand and support the economy of cities. Circular economy actions can support the reduction of the economic gap between linear metabolism and circular metabolism of a city. Gothenburg aims to improve its food sector by implementing circular action across its food value chain. The thesis analyses the food metabolism of Gothenburg, identify suitable circular strategies relevant to the city based on the circularity pillars and model how the selected strategies affect the metabolism change in the production, consumption, trade and waste stages of the food chain. The results of the three modelled actions shows that import dependency on identified food products could be reduced although in some cases dependency shifted from one product to another. The actions could also support more circular food flows in Gothenburg through increased local production, changes in consumption and redistribution of surplus food. Through action alley farming, oat bread production reduced bread imports by 18 percent resulting in additional wheat imports. Action dietary change reduced modelled school meat demand but also increased imports of identified plant-based products. In centralized retail surplus redistribution, the estimated 6 percent increase in surplus capture rate at redistribution centres directed more food to people, that could potentially reduce the amount sent to waste treatment, while also reducing the import dependency of redistributed food products. The results indicate that the selected actions could contribute to more circular food flow in Gothenburg even though trade dependency is not eliminated. The selected actions show potential to support substitution, extend the use of surplus retail food and improve the resilience of the food system.
How Much Information Is Needed for Reliable Groundwater Level Imputation? Evidence from Drought Conditions in Sweden
(2026) Zhang, Yang
Reconstructing groundwater levels during drought is important for assessing water availability. This study investigated the information required for groundwater-level imputation under drought conditions in Sweden. Daily records from 297 monitoring wells were analysed, with drought periods identified using the Standardized Groundwater Index (SGI) and removed to create artificial gaps. Predictors included nearby groundwater-level records, rolling sums of net precipitation (NP), and time variables representing day of year, month and continuous time. Two methods, missForest and ARCHI, were evaluated using RMSE, MAE and Pearson correlation. Configuration comparisons used 147 eligible wells for missForest and 210 for ARCHI, with a separate comparison on 140 common wells. Nearby groundwater-level records were the most important source of information for missForest. NP and time variables provided complementary benefits when reference-well information was limited. Within the eligible sample, approximately 3 to 5 reference wells were sufficient to achieve near-optimal performance, depending on the metric and tolerance, with limited further improvements beyond about 8. Near-optimal performance was defined using 5% and 10% tolerances relative to the best tested configuration. Pearson correlation metric satisfied these criteria with fewer reference wells. On the 140 common wells, missForest W5 (Top10 + NP1–3 + Time) achieved 35.9% lower median RMSE and 39.6% lower median MAE than ARCHI n8, together with higher median correlation. Overall, combining reference-well observations with climatic and time information improved reconstruction performance. Minimum information requirements remained dependent on the evaluation sample, metric and tolerance definition. The descriptive SGI analysis characterised drought variability, while preservation of drought-event characteristics after imputation remains to be evaluated.
Humanoid Balance and Step Recovery in Public Transport Scenarios using Reinforcement Learning
(2026) Shi, Ruochen; Peng, Wei
Humanoid robots that travel by public transport must remain standing while the vehicle floor accelerates, brakes, and turns beneath them. This thesis develops reinforcement-learning policies that keep a 29-degree-of-freedom Unitree G1 balanced on a finite moving floor, using only the robot’s own inertial and joint measurements at deployment. The policies act directly on joint-position targets and are trained with proximal policy optimization and an asymmetric critic in Isaac Lab. Two support conditions form the main study. Without a handle, event-gated rewards allow the robot to step when its current stance can no longer arrest the motion. With a rigid handle, the reward keeps the feet as the primary support, limits the load on the hand, and returns the robot to a support-ready stance after a disturbance. A ceiling-pivoted suspended handle is studied in simulation as an extension. In simulation, the no-handle policy completes almost all straight-line but only about half of the turning episodes of its training distribution, and it recovers by stepping with the disturbance. In a straight-line severity sweep, it leaves the usable floor area from a peak acceleration of about 2 m/s2, whereas the rigid-handle policy does not step and exceeds the 10 N m hand-moment limit of its training from about 4 m/s2. A sweep of the exported policies in MuJoCo shows the same limits. In a moving research van on a closed proving ground, both policies were compared with the manufacturer’s standing controller. Across five paired straight-road maneuvers, the no-handle policy reduced the median torso-attitude variation from 6.6◦ to 1.9◦ and the median whole-body state deviation from 4.7 to 1.6; in turning, it reduced the attitude variation from 3.1◦ to 1.4◦ at the cost of more whole-body motion. The default controller lost balance in S4 and S5 (target speeds of 36 and 45 km/h), leaving 3/5 trials (60%) without tether intervention, compared with 5/5 (100%) for each learned configuration. The rigid-handle configuration produced the smallest median whole-body state deviation in both straight driving (0.039 versus 4.7 for the default condition) and turning (0.131 versus 0.95), with median torso-attitude variations of 1.3◦ and 1.7◦, respectively. The results show that joint-level policies trained on a moving platform transfer to a real vehicle, and that the available support area or the admissible hand load, rather than upright balance itself, sets the limit of the tolerable disturbance.
Bifurcations of Invariant Closed Curves in the Piecewise-Linear 3D Border Collision Normal Form
(2026) Görner, Sven
Invariant closed curves (ICCs) have received increasing attention in the theory of smooth dynamical systems in recent years, with Neimark–Sacker bifurcations being one of the most prominent mechanisms through which they are created. This thesis investigates whether the theory developed for such invariant curves in the smooth setting can be transferred to piecewise-linear systems, and, where it cannot, identifies the mechanisms responsible for the resulting differences. The investigation is carried out for the piecewise-linear three-dimensional Border Collision Normal Form (BCNF), a piecewise-affine map that serves as a canonical local model for systems exhibiting border collision bifurcations. Within this framework, an attracting resonant ICC is understood as being composed of a saddle cycle, a stable node or focus cycle, and the connecting unstable one-dimensional set of the saddle. Building on this notion, the thesis focuses on the three elementary bifurcations associated with eigenvalues leaving the unit circle under parameter perturbation. The thesis examines for each possibility – leaving the unit circle via "+1", "-1" or as a complex conjugated pair anywhere through the unit circle – how an ICC responds when its constituent cycles cross such boundaries in close succession, comparing the resulting behavior to the corresponding smooth bifurcation scenario in order to identify both parallels and fundamental differences. A key difference is that all (non-border-collision) stability-related bifurcations in the BCNF are inherently degenerate; consequently, the stability of newly created invariant sets is governed by a global, border-collision-induced mechanism rather than by a local, smooth-theory-style criterion. As a byproduct of this investigation, a broad numerical exploration of a slice of the BCNF’s parameter space is conducted to catalogue the wide variety of geometric shapes that ICCs can take on in the nonsmooth setting – ranging from needleand star-shaped curves to tightly wound spirals – and to trace these distinctive shapes back to the underlying eigenvalue structure of the constituent cycles and their proximity to the switching manifold. The results are obtained using the numerical tools PeriodScanner, for the detection and stability classification of periodic orbits, and Manifis, for the high-precision computation of their one-dimensional invariant sets of their saddle cycles.
Exploring the Benefits of Compression to Improve Data-intensive Neural Network Inference
(2026) Meidinger, Max Andreas Peter
Convolutional Neural Networks (CNNs) have become essential for many computer vision tasks such as image classification, object detection and semantic segmentation. State-of-the-art models come with high computational demands that require efficient acceleration. FPGAs are a popular choice, as they can host energy-efficient, high-throughput inference designs. Streaming dataflow designs are particularly efficient, as they create custom per-layer pipelines which confine most of the data movement on chip. As these accelerators also prefer to keep model parameters on chip, on-chip buffer space becomes a limiting factor. Long skip connections, a ubiquitous structure of today’s networks, aggravate this problem further, as they require large amounts of buffer space to synchronize the diverging branches of the model’s architecture. Off-chip eviction instead moves the activations of skip connections temporarily to off-chip memory, but this puts additional load on the memory bus, which can lead to stalls in the compute pipeline if the processing engines cannot be fed. This problem can be alleviated using lossless compression, but lossless compression is fundamentally limited. This thesis explores lossy compression of the evicted activations instead, trading a bounded loss of inference accuracy for a higher compression ratio, without retraining. Starting from SZ1, a prediction-based compressor whose model fits the streaming paradigm, we develop ASZ (Activation SZ) in a design space exploration, which adds block-level error thresholds, new prediction functions, per-channel selection of the prediction functions, and Huffman coding and downsampling stages. We evaluate ASZ against lossless (Huffman coding, RLE) and lossy baselines (quantization with Huffman coding (QH), Lorenzo prediction, SZ1, downsampling, SVD and JPEG) on the 74 skip connections of eight CNN classifiers on ImageNet, CIFAR-10 and CIFAR-100, and estimate the hardware cost of every method. Lossy compression goes well beyond lossless encoding: on every connection, a lossy method exceeds the best lossless one, on 71 of them at a cost of at most 0 1 pp of top-1 accuracy. Within a budget of 1 pp, the streaming-compatible compressors QH, Lorenzo and ASZ reach geometric-mean compression ratios of 3 9, 3 8 and 3 5, 2 3 to 2 6 times the 1 5 of Huffman coding, while RLE does not compress on average. ASZ improves on SZ1 (2 7), but trails the quantizers on most connections, since SZ1-type compressors spend their bits on the choice of a prediction function rather than on the values or their residuals. JPEG reaches 6 6, but buffers eight rows per channel and needs multipliers on the data path. When several connections are compressed, their errors add up and the ratio within the budget falls, least for the error-bounded Lorenzo and SZ1; with three connections, the best lossy method still stays 1 5 to 4 3 times above Huffman coding. Lossy compression can thus substantially reduce the off-chip traffic of evicted skip connections, and QH, the cheapest lossy compressor and the best streaming-compatible one in ratio, offers the best trade-off between compression ratio and hardware cost.