| 1. |
- Kleyko, Denis, et al.
(författare)
-
Vehicle Classification using Road Side Sensors and Feature-free Data Smashing Approach
- 2016
-
Ingår i: 2016 IEEE 19th International Conference on Intelligent Transportation Systems (ITSC). - Piscataway : IEEE. - 9781509018895 - 9781509018888 - 9781509018901 ; , s. 1988-1993
-
Konferensbidrag (refereegranskat)abstract
- The main contribution of this paper is a study of the applicability of data smashing - a recently proposed data mining method - for vehicle classification according to the "Nordic system for intelligent classification of vehicles" standard, using measurements of road surface vibrations and magnetic field disturbances caused by passing vehicles. The main advantage of the studied classification approach is that it, in contrast to the most of traditional machine learning algorithms, does not require the extraction of features from raw signals. The proposed classification approach was evaluated on a large dataset consisting of signals from 3074 vehicles. Hence, a good estimate of the actual classification rate was obtained. The performance was compared to the previously reported results on the same problem for logistic regression. Our results show the potential trade-off between classification accuracy and classification method's development efforts could be achieved.
|
|
| 2. |
- Rezk, Nesma, 1987-, et al.
(författare)
-
MOHAQ : Multi-Objective Hardware-Aware Quantization of recurrent neural networks
- 2022
-
Ingår i: Journal of systems architecture. - Amsterdam : Elsevier BV. - 1383-7621 .- 1873-6165. ; 133
-
Tidskriftsartikel (refereegranskat)abstract
- The compression of deep learning models is of fundamental importance in deploying such models to edge devices. The selection of compression parameters can be automated to meet changes in the hardware platform and application. This article introduces a Multi-Objective Hardware-Aware Quantization (MOHAQ) method, which considers hardware performance and inference error as objectives for mixed-precision quantization. The proposed method feasibly evaluates candidate solutions in a large search space by relying on two steps. First, post-training quantization is applied for fast solution evaluation (inference-only search). Second, we propose the "beacon-based search" to retrain selected solutions only and use them as beacons to estimate the effect of retraining on other solutions. We use speech recognition models on TIMIT dataset. Experimental evaluations show that Simple Recurrent Unit (SRU)-based models can be compressed up to 8x by post-training quantization without any significant error increase. On SiLago, we found solutions that achieve 97% and 86% of the maximum possible speedup and energy saving, with a minor increase in error on an SRU-based model. On Bitfusion, the beacon-based search reduced the error gain of the inference-only search on SRU-based models and Light Gated Recurrent Unit (LiGRU)-based model by up to 4.9 and 3.9 percentage points, respectively.
|
|
| 3. |
- Sandberg, Karin, et al.
(författare)
-
Renovation of an office building with prefabricated wooden element – case Hedensbyn
- 2018
-
Ingår i: Cold Climate HVAC 2018. - Cham : Springer. - 9783030006617 - 9783030006624 ; , s. 587-598
-
Konferensbidrag (refereegranskat)abstract
- There is a major need of cost-effective renovation that leads to lower energy consumption and better environment. This article shows the results from a pilot case of a newly developed prefabricatedbuilding system. It is an industrially prefabricated insulated wooden element adapted to renovationand upgrading of building envelopes. The renovated building is a one-story office building located in Skellefted in the north of Sweden. Energy performance, thermal bridges, risk of moisture problems, LCA, applicability of the renovation method and assembly time were evaluated during the planning and execution of the renovation. Results from this case show that the elements were very light and easy for one person to handle at the building site. There is a great potential to reduce assembly time with improved joints and element sizes adapted to the building as well as improved batch packaging from the factory. With 100 mm insulation, the renovation gives a certain energy savings, and LCA calculations show that the reduction of climate impact due to reduced heating energy used during a service life 50 years corresponds to the climate impact of the renovation measures. The risk ofmicrobial growth can be regarded as small.
|
|
| 4. |
- Savas, Süleyman, 1986-, et al.
(författare)
-
A framework to generate domain-specific manycore architectures from dataflow programs
- 2020
-
Ingår i: Microprocessors and microsystems. - Amsterdam : Elsevier. - 0141-9331 .- 1872-9436. ; 72
-
Tidskriftsartikel (refereegranskat)abstract
- In the last 15 years we have seen, as a response to power and thermal limits for current chip technologies, an explosion in the use of multiple and even many computer cores on a single chip. But now, to further improve performance and energy efficiency, when there are potentially hundreds of computing cores on a chip, we see a need for a specialization of individual cores and the development of heterogeneous manycore computer architectures.However, developing such heterogeneous architectures is a significant challenge. Therefore, we propose a design method to generate domain specific manycore architectures based on RISC-V instruction set architecture and automate the main steps of this method with software tools. The design method allows generation of manycore architectures with different configurations including core augmentation through instruction extensions and custom accelerators. The method starts from developing applications in a high-level dataflow language and ends by generating synthesizable Verilog code and cycle accurate emulator for the generated architecture.We evaluate the design method and the software tools by generating several architectures specialized for two different applications and measure their performance and hardware resource usages. Our results show that the design method can be used to generate specialized manycore architectures targeting applications from different domains. The specialized architectures show at least 3 to 4 times better performance than the general purpose counterparts. In certain cases, replacing general purpose components with specialized components saves hardware resources. Automating the method increases the speed of architecture development and facilitates the design space exploration of manycore architectures.
|
|
| 5. |
- Soleimani-Mohseni, Mohsen, et al.
(författare)
-
Energy simulation for a high-rise building using IDA ICE : investigations in different climates
- 2016
-
Ingår i: Building Simulation. - Beijing : Tsinghua University Press. - 1996-3599 .- 1996-8744. ; 9:6, s. 629-640
-
Tidskriftsartikel (refereegranskat)abstract
- In this paper a model of a high-rise building is constructed in the simulation program IDA ICE. The model is based on an IFC-model of a demonstration building constructed in Ljubljana, Slovenia, as part of an EU-project, EE-high-rise. The model's energy performance was simulated for four cities: Umeå (Scandinavia), Ljubljana (Central Europe), Sibenik (Mediterranean) and Dubai (The Persian Gulf). Furthermore, the climate envelope of the building was modified with the aim to improve the model's energy performance in each of the regions. The results were evaluated according to the energy requirements of passive house standard by the German Passive House Institute. The analysis suggests that the reference building model, which itself incorporates several energy efficient components, was unable to meet the German passive house standard in none of the four cities (Umeå, Ljubljana, Sibenik and Dubai) studied. By providing a combination of energy saving measures, such as modifications of thermal resistance of building envelope, the building may be able to meet the passive house standard in Ljubljana. The analysis concludes that the reduction in window area results in reduction of both heating and cooling demand. Increase in the thickness of the insulation and the thermal resistance of windows reduces the space heating demand for Umeå, Ljubljana and Sibenik (not applied for Dubai) while increasing the cooling demand for these cities. Increased airtightness has marginal effect on heating and cooling demand for all investigated cities. Reduced thermal resistance of windows will decrease cooling demand for Ljubljana, Sibenik and Dubai (not applied for Umeå). Reduced insulation thickness (not applied for Umeå) will decrease cooling demand for Ljubljana and Sibenik but not for Dubai. Reducing the insulation thickness may often result in reduced cooling demand for moderately warm countries since the average outdoor temperature could be lower than the indoor temperature during part of the cooling season. In those situations a reduced insulation thickness can cause heat flow from the relatively hot inside to the colder outside. However, for hot climates like in Dubai where outdoor temperature is higher than the indoor temperature for most of the year, reducing the insulation thickness will increase the cooling demand. This result suggests that the insulation thickness must be chosen and optimized based on heating and cooling demand, internal heat gain, and outdoor climate
|
|
| 6. |
- Brembilla, Christian, 1983-, et al.
(författare)
-
Investigations of thermal parameters addressed to a building simulation model
- 2015
-
Ingår i: Proceedings of BS2015. - India, Hyderabad : International Building Performance Simulation Association (IBPSA). ; , s. 2741-2748, s. 2741-2748
-
Konferensbidrag (refereegranskat)abstract
- This paper shows the tolerance of thermal parameters addressed to a building simulation model in relation to the local control of the HVAC system. This work is suitable for a modeler that has to set up a building simulation model. The modeler has to know which parameter needs to be considered carefully and vice-versa which does not need deep investigations. Local differential sensitivity analysis of thermal parameters generates the uncertainty bands for the indoor air. The latter operation is repeated with P, PI and PID local control of the heating system. In conclusion, the local control of a room has a deterministic impact on the tolerance of thermal parameters.
|
|
| 7. |
- Brembilla, Christian, 1983-, et al.
(författare)
-
Transient model of a panel radiator
- 2015
-
Ingår i: Energy, Science and Technology 2015. - Karlsruher, Germany : Karlsruher Institut für Technologie (KIT). ; , s. 321-321, s. 321-321
-
Konferensbidrag (refereegranskat)abstract
- This paper shows a detailed transient model of a panel radiator considered as a system of multiple storageelements. The experiment records the temperature surface of the panel in the process of heating up. Thequalitative results of the experiment suggest the more appropriate technique for modelling this technology. The transient model performs the modelling with horizontal thermal capacitances connected in series. This modelcalculates the temperature of exhaust flow, heat emission towards indoor environment, temperature gradient onpanel surface, dead and balancing time identified numerically on the chart.
|
|
| 8. |
- Rezk, Nesma, 1987-, et al.
(författare)
-
Recurrent Neural Networks : An Embedded Computing Perspective
- 2020
-
Ingår i: IEEE Access. - Piscataway : IEEE. - 2169-3536. ; 81:1, s. 57967-57996
-
Tidskriftsartikel (refereegranskat)abstract
- Recurrent Neural Networks (RNNs) are a class of machine learning algorithms used for applications with time-series and sequential data. Recently, there has been a strong interest in executing RNNs on embedded devices. However, difficulties have arisen because RNN requires high computational capability and a large memory space. In this paper, we review existing implementations of RNN models on embedded platforms and discuss the methods adopted to overcome the limitations of embedded systems. We will define the objectives of mapping RNN algorithms on embedded platforms and the challenges facing their realization. Then, we explain the components of RNN models from an implementation perspective. We also discuss the optimizations applied to RNNs to run efficiently on embedded platforms. Finally, we compare the defined objectives with the implementations and highlight some open research questions and aspects currently not addressed for embedded RNNs. Overall, applying algorithmic optimizations to RNN models and decreasing the memory access overhead is vital to obtain high efficiency. To further increase the implementation efficiency, we point up the more promising optimizations that could be applied in future research. Additionally, this article observes that high performance has been targeted by many implementations, while flexibility has, as yet, been attempted less often. Thus, the article provides some guidelines for RNN hardware designers to support flexibility in a better manner.
|
|
| 9. |
- Kadi, F., et al.
(författare)
-
The effects of different training programs on the trapezius muscle of women with work-related neck and shoulder myalgia
- 2000
-
Ingår i: Acta Neuropathologica. - Berlin : Springer Berlin/Heidelberg. - 0001-6322 .- 1432-0533. ; 100:3, s. 253-258
-
Tidskriftsartikel (refereegranskat)abstract
- The aim of this study was to examine the effects of training on the structural characteristics of the trapezius muscle in women with work-related trapezius myalgia. Muscle biopsies were taken before and after 10 weeks of three different training programs (strength, endurance and coordination). Enzyme-immunohistochemical analysis was performed to assess muscle fibre types, fibre area, capillary supply and cytochrome c oxidase (COX) activity. There was an increase in the proportion of type LIA fibres in strength trained group (P < 0.05). Strength training elicited a preferential increase in the area of type II fibres (P < 0.05); both strength and endurance programs induced an increase in the number of capillaries around type I and IIA muscle fibres. Finally, all training programs induced a decrease in the proportion of COX-negative fibres. In conclusion, the trapezius muscle of women with neck and shoulder myalgia is characterised by a great potential of adaptation to physical exercise over a period of 10 weeks. The significant changes in the number of capillaries and the specific changes induced by training at the level of muscle fibres might well explain the improvement of muscle function.
|
|
| 10. |
- Brembilla, Christian, 1983-, et al.
(författare)
-
Paradoxes in understanding the Efficiency Factors of Space Heating
- 2019
-
Ingår i: Energy Efficiency. - Dordrecht : Springer. - 1570-646X .- 1570-6478. ; 12:3, s. 777-786
-
Tidskriftsartikel (refereegranskat)abstract
- Efficiency factors are here defined as the thermal energy performance indicators of the space heating. Until recently, the efficiency factors were assumed as one value for space heating located in any climate. This study addresses the problem of how the outdoor climate affects the efficiency factors of a space heating equipped with 1D model of hydronic floor heating. The findings show how the efficiency factors, computed with two numerical methods, are correlated with the solar radiation. This study highlights the paradoxes in understanding the results of efficiency factors analysis. This work suggests how to interpret and use the efficiency factors as a benchmark performance indicator.
|
|
