Publications from the Department of Electrical Engineering
Automatic ControlCommunication Systems
Computer Engineering
Computer Vision
Information Coding
Integrated Circuits and Systems
Vehicular Systems
Latest PhD theses
New transport technologies have the potential to create more efficient modes of transport and transforming cities for the better by improving urban productivity and increasing efficiency of its transport system to move consumers, labor, and freight. Traffic accidents, energy consumption, pollution, congestion, and long commuting times are main concerns and new transport technologies with autonomous vehicles have the potential to be part of the solution to these important challenges.
An autonomous, or highly automated car is a vehicle that can operate with little to no human assistance. This technology is not yet generally available, but if fully realized have the potential to fundamentally change the transportation system. The passenger experience will fundamentally change, but there are also possibilities to increase traffic flow, form platoons of transport vehicles to reduce air-drag and thereby energy consumption, and a main challenge is to realize all this in a safe way in uncertain and complex traffic situations on highways and in urban scenarios.
The key topic of this dissertation is how optimal control techniques, more specifically Model Predictive Control (MPC), can be applied in autonomous driving in dynamic environments and with dynamic constraints on vehicle behavior. The main problem studied is how to control multiple vehicles in an optimal, safe, and collision free way in complex traffic scenarios, e.g., laneswitching, merging, or intersection situations in the presence of moving obstacles, i.e., other vehicles whose behavior and intent may not be known. Further, the controller needs to take maneuvering capabilities of the vehicle into account, respecting road boundaries, speed limitations, and other traffic rules. Optimization-based techniques for control are interesting candidates for multi-vehicle problems, respecting well-defined rules in traffic while still providing a high degree of decision autonomy to each vehicle.
To ensure autonomy, it is studied how to decentralize the control approach to not rely on a centralized computational resource. Different methods and approaches are proposed in the thesis with guaranteed convergence and collision-avoidance features. To reduce the computational complexity of the controller, a Gaussian risk model for collision prediction is integrated and also a technique that combines MPC with learning methods is explored.
Main contributions of this dissertation are control methods for autonomous vehicles that provide safety and comfort of passengers even in uncertain traffic situations where the behavior of surrounding vehicles is uncertain, and the methods are computationally fast enough to be used in real time. An important property is that the proposed algorithms are general enough to be used in different traffic scenarios, hence reducing the need for specific solutions for specific situations.
@phdthesis{diva2:1507702,
author = {Mohseni, Fatemeh},
title = {{Decentralized Optimal Control for Multiple Autonomous Vehicles in Traffic Scenarios}},
school = {Linköping University},
type = {{Linköping Studies in Science and Technology. Dissertations No. 2116}},
year = {2021},
address = {Sweden},
}
Without a driver to fall back on, a fully self-driving car needs to be able to handle any situation it can encounter. With the perspective of future safety systems, this research studies autonomous maneuvering at the tire-road friction limit. In these situations, the dynamics is highly nonlinear, and the tire-road parameters are uncertain.
To gain insights into the optimal behavior of autonomous safety-critical maneuvers, they are analyzed using optimal control. Since analytical solutions of the studied optimal control problems are intractable, they are solved numerically. An optimization formulation reveals how the optimal behavior is influenced by the total amount of braking. By studying how the optimal trajectory relates to the attainable forces throughout a maneuver, it is found that maximizing the force in a certain direction is important. This is like the analytical solutions obtained for friction-limited particle models in earlier research, and it is shown to result in vehicle behavior close to the optimal also for a more complex model.
Based on the insights gained from the optimal behavior, controllers for autonomous safety maneuvers are developed. These controllers are based on using acceleration-vector references obtained from friction-limited particle models. Exploiting that the individual tire forces tend to be close to their friction limits, the desired tire slip angles are determined for a given acceleration-vector reference. This results in controllers capable of operating at the limit of friction at a low computational cost and reduces the number of vehicle parameters used. For straight-line braking, ABS can intervene to reduce the braking distance without prior information about the road friction. Inspired by this, a controller that uses the available actuation according to the least friction necessary to avoid a collision is developed, resulting in autonomous collision avoidance without any estimation of the tire–road friction.
Investigating time-optimal lane changes, it is found that a simple friction-limited particle model is insufficient to determine the desired acceleration vector, but including a jerk limit to account for the yaw dynamics is sufficient. To enable a tradeoff between braking and avoidance with a more general obstacle representation, the acceleration-vector reference is computed in a receding-horizon framework.
The controllers developed in this thesis show great promise with low computational cost and performance not far from that obtained offline by using numerical optimization when evaluated in high-fidelity simulation.
@phdthesis{diva2:1478641,
author = {Fors, Victor},
title = {{Autonomous Vehicle Maneuvering at the Limit of Friction}},
school = {Linköping University},
type = {{Linköping Studies in Science and Technology. Dissertations No. 2102}},
year = {2020},
address = {Sweden},
}
The fifth generation of mobile communication systems (5G) is nowadays a reality. 5G networks are been deployed all over the world, and the first 5G-capable devices (e.g., smartphones, tablets, wearable, etc.) are already commercially available. 5G systems provide unprecedented levels of connectivity and quality of service (QoS) to cope with the incessant growth in the number of connected devices and the huge increase in data-rate demand.
Massive MIMO (multiple-input multiple-output) technology plays a key role in 5G systems. The underlying principle of this technology is the use of a large number of co-located antennas at the base station, which coherently transmit/receive signals to/from multiple users. This signal co-processing at multiple antennas leads to manifold benefits: array gain, spatial diversity and spatial user multiplexing. These elements enable to meet the QoS requirements established for the 5G systems. The major bottleneck of massive MIMO systems as well as of any cellular network is the inter-cell interference, which affects significantly the cell-edge users, whose performance is already degraded by the path attenuation. To overcome these limitations and provide uniformly excellent service to all the users we need a more radical approach: we need to challenge the cellular paradigm.
In this regard, cell-free massive MIMO constitutes the paradigm shift. In the cell-free paradigm, it is not the base station surrounded by the users, but rather it is each user being surrounded by smaller, simpler, serving base stations referred to as access points (APs). In such a system, each user experiences being in the cell-center, and it does not experience any cell boundaries. Hence, the terminology cell-free. As a result, users are not affected by inter-cell interference, and the path attenuation is significantly reduced due to the presence of many APs in their proximity. This leads to impressive performance.
Although appealing from the performance viewpoint, the designing and implementation of such a distributed massive MIMO system is a challenging task, and it is the object of this thesis. More specifically, in this thesis we study:
Paper A) The large potential of this promising technology in realistic indoor/outdoor scenarios while also addressing practical deployment issues, such as clock synchronization among APs, and cost-efficient implementations. We provide an extensive description of a cell-free massive MIMO system, emphasizing strengths and weaknesses, and pointing out differences and similarities with existing distributed multiple antenna systems, such as Coordinated MultiPoint (CoMP).
Paper B) How to preserve the scalability of the system, by proposing a solution related to data processing, network topology and power control. We consider a realistic scenario where multiple central processing units serve disjoint subsets of APs, and compare the spectral efficiency provided by the proposed scalable framework with the canonical cell-free massive MIMO and CoMP.
Paper C) How to improve the spectral efficiency (SE) in the downlink (DL), by devising two distributed precoding schemes, referred to as local partial zero-forcing (ZF) and local protective partial ZF, that provide an adaptable trade-off between interference cancelation and boosting of the desired signal, with no additional front-haul overhead, and that are implementable by APs with very few antennas. We derive closed-form expressions for the achievable SE under the assumption of independent Rayleigh fading channel, channel estimation error and pilot contamination. These closed-form expressions are then used to devise optimal max-min fairness power control.
Paper D) How to further improve the SE by letting the user estimate the DL channel from DL pilots, instead of relying solely on the knowledge of the channel statistics. We derive an approximate closed-form expression of the DL SE for conjugate beamforming (CB), and assuming independent Rayleigh fading. This expression accounts for beamformed DL pilots, estimation errors and pilot contamination at both the AP and the user side. We devise a sequential convex approximation algorithm to globally solve the max-min fairness power control optimization problem, and a greedy algorithm for uplink (UL) and DL pilot assignment. The latter consists in jointly selecting the UL and DL pilot pair, for each user, that maximizes the smallest SE in the network.
Paper E) A precoding scheme that is more suitable when only the channel statistics are available at the users, referred to as enhanced normalized CB. It consists in normalizing the precoding vector by its squared norm in order to reduce the fluctuations of the effective channel seen at the user, and thereby to boost the channel hardening. The performance achieved by this scheme is compared with the CB scheme with DL training (described in Paper D).
Paper F) A maximum-likelihood-based method to estimate the channel statistics in the UL, along with an accompanying pilot transmission scheme, that is particularly useful in line-of-sight operation and in scenarios with resource constraints. Pilots are structurally phase-rotated over different coherence blocks to create an effective statistical distribution of the received pilot signal that can be efficiently exploited by the AP when performing the proposed estimation method.
The overall conclusion is that cell-free massive MIMO is not a utopia, and a practical, distributed, scalable, high-performance system can be implemented. Today it represents a hot research topic, but tomorrow it might represent a key enabler for beyond-5G technology, as massive MIMO has been for 5G.
@phdthesis{diva2:1448945,
author = {Interdonato, Giovanni},
title = {{Cell-Free Massive MIMO:
Scalability, Signal Processing and Power Control}},
school = {Linköping University},
type = {{Linköping Studies in Science and Technology. Dissertations No. 2090}},
year = {2020},
address = {Sweden},
}
The control-theoretic notion of controllability captures the ability to guide a system toward a desired state with a suitable choice of inputs. Controllability of complex networks such as traffic networks, gene regulatory networks, power grids etc. can for instance enable efficient operation or entirely new applicative possibilities. However, when control theory is applied to complex networks like these, several challenges arise. This thesis considers some of them, in particular we investigate how a given network can be rendered controllable at a minimum cost by placement of control inputs or by growing the network with additional edges between its nodes. As cost function we take either the number of control inputs that are needed or the energy that they must exert.
A control input is called unilateral if it can assume either positive or negative values, but not both. Motivated by the many applications where unilateral controls are common, we reformulate classical controllability results for this particular case into a more computationally-efficient form that enables a large scale analysis. Assuming that each control input targets only one node (called a driver node), we show that the unilateral controllability problem is to a high degree structural: from topological properties of the network we derive theoretical lower bounds for the minimal number of unilateral control inputs, bounds similar to those that have already been established for the minimal number of unconstrained control inputs (e.g. can assume both positive and negative values). With a constructive algorithm for unilateral control input placement we also show that the theoretical bounds can often be achieved.
A network may be controllable in theory but not in practice if for instance unreasonable amounts of control energy are required to steer it in some direction. For the case with unconstrained control inputs, we show that the control energy depends on the time constants of the modes of the network, the longer they are, the less energy is required for control. We also present different strategies for the problem of placing driver nodes such that the control energy requirements are reduced (assuming that theoretical controllability is not an issue). For the most general class of networks we consider, directed networks with arbitrary eigenvalues (and thereby arbitrary time constants), we suggest strategies based on a novel characterization of network non-normality as imbalance in the distribution of energy over the network. Our formulation allows to quantify network non-normality at a node level as combination of two different centrality metrics. The first measure quantifies the influence that each node has on the rest of the network, while the second measure instead describes the ability to control a node indirectly from the other nodes. Selecting the nodes that maximize the network non-normality as driver nodes significantly reduces the energy needed for control.
Growing a network, i.e. adding more edges to it, is a promising alternative to reduce the energy needed to control it. We approach this by deriving a sensitivity function that enables to quantify the impact of an edge modification with the H2 and H∞ norms, which in turn can be used to design edge additions that improve commonly used control energy metrics.
@phdthesis{diva2:1425446,
author = {Lindmark, Gustav},
title = {{Controllability of Complex Networks at Minimum Cost}},
school = {Linköping University},
type = {{Linköping Studies in Science and Technology. Dissertations No. 2074}},
year = {2020},
address = {Sweden},
}
During the last decades, improved sensor and hardware technologies as well as new methods and algorithms have made self-driving vehicles a realistic possibility in the near future. At the same time, there has been a growing demand within the transportation sector to increase efficiency and to reduce the environmental impact related to transportation of people and goods. Therefore, many leading automotive and technology companies have turned their attention towards developing advanced driver assistance systems and self-driving vehicles.
Autonomous vehicles are expected to have their first big impact in closed environments, such as mines, harbors, loading and offloading sites. In such areas, the legal requirements are less restrictive and the surrounding environment is more controlled and predictable compared to urban areas. Expected positive outcomes include increased productivity and safety, reduced emissions and the possibility to relieve the human from performing complex or dangerous tasks. Within these sites, tractor-trailer vehicles are frequently used for transportation. These vehicles are composed of several interconnected vehicle segments, and are therefore large, complex and unstable while reversing. This thesis addresses the problem of designing efficient motion planning and feedback control techniques for such systems.
The contributions of this thesis are within the area of motion planning and feedback control for long tractor-trailer combinations operating at low-speeds in closed and unstructured environments. It includes development of motion planning and feedback control frameworks, structured design tools for guaranteeing closed-loop stability and experimental validation of the proposed solutions through simulations, lab and field experiments. Even though the primary application in this work is tractor-trailer vehicles, many of the proposed approaches can with some adjustments also be used for other systems, such as drones and ships.
The developed sampling-based motion planning algorithms are based upon the probabilistic closed-loop rapidly exploring random tree (CL-RRT) algorithm and the deterministic lattice-based motion planning algorithm. It is also proposed to use numerical optimal control offline for precomputing libraries of optimized maneuvers as well as during online planning in the form of a warm-started optimization step.
To follow the motion plan, several predictive path-following control approaches are proposed with different computational complexity and performance. Common for these approaches are that they use a path-following error model of the vehicle for future predictions and are tailored to operate in series with a motion planner that computes feasible paths. The design strategies for the path-following approaches include linear quadratic (LQ) control and several advanced model predictive control (MPC) techniques to account for physical and sensing limitations. To strengthen the practical value of the developed techniques, several of the proposed approaches have been implemented and successfully demonstrated in field experiments on a full-scale test platform. To estimate the vehicle states needed for control, a novel nonlinear observer is evaluated on the full-scale test vehicle. It is designed to only utilize information from sensors that are mounted on the tractor, making the system independent of any sensor mounted on the trailer.
@phdthesis{diva2:1424832,
author = {Ljungqvist, Oskar},
title = {{Motion planning and feedback control techniques with applications to long tractor-trailer vehicles}},
school = {Linköping University},
type = {{Linköping Studies in Science and Technology. Dissertations No. 2070}},
year = {2020},
address = {Sweden},
}
Trilateration is the mathematical theory of computing the intersection of circles. These circles may be obtained by time of flight (ToF) measurements in radio systems, as well as laser, radar and sonar systems. A first purpose of this thesis is to survey recent efforts in the area and their potential for localization. The rest of the thesis then concerns selected problems in new cellular radio standards as well as fundamental challenges caused by propagation delays in the ToF measurements, which cannot travel faster than the speed of light. We denote the measurement uncertainty stemming from propagation delays for positive noise, and develop a general theory with optimal estimators for selected distributions, which can be applied to trilateration but also a much wider class of estimation problems.
The first contribution concerns a narrow-band mode in the long-term evolution (LTE) standard intended for internet of things (IoT) devices. This LTE standard includes a special position reference signal sent synchronized by all base stations (BS) to all IoT devices. Each device can then compute several pair-wise time differences that correspond to hyperbolic functions. The simulation-based performance evaluation indicates that decent position accuracy can be achieved despite the narrow bandwidth of the channel.
The second contribution is a study of how timing measurements in LTE can be combined. Round trip time (RTT) to the serving BS and time difference of arrival (TDOA) to the neighboring BS are used as measurements. We propose a filtering framework to deal with the existing uncertainty in the solution and evaluate with both simulated and experimental test data. The results indicate that the position accuracy is better than 40 meters 95% of the time.
The third contribution is a comprehensive theory of how to estimate the signal observed in positive noise, that is, random variables with positive support. It is well known from the literature that order statistics give one order of magnitude lower estimation variance compared to the best linear unbiased estimator (BLUE). We provide a systematic survey of some common distributions with positive support, and provide derivations and summaries of estimators based on order statistics, including the BLUE one for comparison. An iterative global navigation satellite system (GNSS) localization algorithm, based on the derived estimators, is introduced to jointly estimate the receiver’s position and clock bias.
The fourth contribution is an extension of the third contribution to a particular approach to utilize positive noise in nonlinear models. That is, order statistics have been employed to derive estimators for a generic nonlinear model with positive noise. The proposed method further enables the estimation of the hyperparameters of the underlying noise distribution. The performance of the proposed estimator is then compared with the maximum likelihood estimator when the underlying noise follows either a uniform or exponential distribution.
@phdthesis{diva2:1393383,
author = {Radnosrati, Kamiar},
title = {{Time of Flight Estimation for Radio Network Positioning}},
school = {Linköping University},
type = {{Linköping Studies in Science and Technology. Dissertations No. 2054}},
year = {2020},
address = {Sweden},
}
The amount of goods produced and transported around the world each year increases and heavy-duty trucks are an important link in the logistic chain. To guarantee reliable delivery a high degree of availability is required, i.e., avoid standing by the road unable to continue the transport mission. Unplanned stops by the road do not only cost due to the delay in delivery, but can also lead to damaged cargo. Vehicle downtime can be reduced by replacing components based on statistics of previous failures. However, such an approach is both expensive due to the required frequent visits to a workshop and inefficient as many components from the vehicles in the fleet are still operational. A prognostic method, allowing for vehicle individualized maintenance plans, therefore poses a significant potential in the automotive field. The prognostic method estimates component degradation and remaining useful life based on recorded data and how the vehicle has been operated.
Lead-acid batteries is a part of the electrical power system in a heavy-duty truck, primarily responsible for powering the starter motor but also powering auxiliary units, e.g., cabin heating and kitchen equipment, which makes the battery a vital component for vehicle availability. Developing physical models of battery degradation is a difficult process which requires access to battery health sensing that is not available in the given study as well a detailed knowledge of battery chemistry.
An alternative approach, considered in this work, is data-driven methods based on large amounts of logged data describing vehicle operation conditions. In the use-case studied, recorded data is not closely related to battery health which makes battery prognostic challenging. Data is collected during infrequent and non-equidistant visits to a workshop and there are complex dependencies between variables in the data. The main aim of this work has been to develop a framework and methods for estimating lifetime of lead-acid batteries using data-driven methods for condition-based maintenance. The methodology is general and can be applicable for prognostics of other components.
A main contribution of the thesis is development of machine learning models for predictive maintenance, estimating conditional reliability functions, using Random Survival Forests (RSF) and recurrent neural networks (RNN). An important property of the data is that for a specific vehicle there may be multiple data readouts, but also one single data readout which makes predictive modeling challenging and dealing with this situation is discussed for both RSF and neural networks models. Data quality is important when building data-driven models, and here the data is imbalanced since there are few battery failures relative to the number of vehicles. Further, the data includes many uninformative variables and among those that are informative, there are complex dependencies and correlation. Methods for selecting which data features to use in the model in this situation is also a key contribution. When a point estimation of the conditional reliability functions is available, it is of interest to know how uncertain the estimate is as it allows to take quality of the prediction into account when deciding on maintenance actions. A theory for estimating the variance of the RSF predictor is another contribution in the thesis. To conclude, the results show that Long Short-Term Memory networks, which is a type of RNN, is the most suitable for the vehicle operational data and give the best performance among methods evaluated in the thesis.
@phdthesis{diva2:1377581,
author = {Voronov, Sergii},
title = {{Machine Learning Models for Predictive Maintenance}},
school = {Linköping University},
type = {{Linköping Studies in Science and Technology. Dissertations No. 2040}},
year = {2020},
address = {Sweden},
}
The development of information and communication technologies (ICT) provides the means for reaching global connectivity that can help humanity progress and prosper. This comes with high demands on data traffic and number of connected devices which are rapidly growing and need to be met by technological development. Massive MIMO, where MIMO stands for multiple-input multiple-output, is a fundamental component of the 5G wireless communication standard for its ability to provide high spectral and energy efficiency, SE and EE, respectively. The key feature of this technology is the use of a large number of antennas at the base stations (BSs) to spatially multiplex several user equipments (UEs).
In the development of new technologies like Massive MIMO, many design alternatives need to be evaluated and compared in order to find the best operating point with a preferable tradeoff between low cost and complexity. In this thesis, two alternative designs for signal processing and hardware in Massive MIMO are studied and compared with the baseline operation in terms of SE, EE, and power consumption. The first design is called superimposed pilot (SP) transmission and is based on superimposing pilot and data symbols to eliminate the need to reserve dedicated time-frequency resources for pilots. This allows more data to be transmitted and supports longer pilot sequences that, in turn, reduce pilot contamination. The second design is mixed analog-to-digital converters (ADCs) and it aims at balancing the SE performance and the power consumption cost by allowing different ADC bit resolutions across the BS antennas.
The results show that the Massive MIMO baseline, when properly optimized, is the preferred choice in standard deployments and propagation conditions. However, the SP alternative design can increase the SE compared to the baseline by using the Massive-MIMO iterative channel estimation and decoding (MICED) algorithm proposed in this dissertation. In particular, the SE gains are found in cases with high mobility, high carrier frequencies, or high number of spatially multiplexed UEs. For the mixed-ADCs alternative design, improvements in the SE and EE compared to the Massive MIMO baseline can be achieved in cases with distributed BS antennas where interference suppression techniques are used.
@phdthesis{diva2:1385646,
author = {Verenzuela, Daniel},
title = {{Exploring Alternative Massive MIMO Designs:
Superimposed Pilots and Mixed-ADCs}},
school = {Linköping University},
type = {{Linköping Studies in Science and Technology. Dissertations No. 2041}},
year = {2020},
address = {Sweden},
}
Massive MIMO (multiple-input multiple-output) is considered as an heir of the multi-user MIMO technology and it has gained lots of attention from both academia and industry since the last decade. By equipping base stations (BSs) with hundreds of antennas in a compact array or a distributed manner, this new technology can provide very large multiplexing gains by serving many users on the same time-frequency resources and thereby bring significant improvements in spectral efficiency (SE) and energy efficiency (EE) over the current wireless networks. The transmit power, pilot training, and spatial transmission resources need to be allocated properly to the users to achieve the highest possible performance. This is called resource allocation and can be formulated as design utility optimization problems. If the resource allocation in Massive MIMO is optimized, the technology can handle the exponential growth in both wireless data traffic and number of wireless devices, which cannot be done by the current cellular network technology.
In this thesis, we focus on the five different resource allocation aspects in Massive MIMO communications: The first part of the thesis studies if power control and advanced coordinated multipoint (CoMP) techniques are able to bring substantial gains to multi-cell Massive MIMO systems compared to the systems without using CoMP. More specifically, we consider a network topology with no cell boundary where the BSs can collaborate to serve the users in the considered coverage area. We focus on a downlink (DL) scenario in which each BS transmits different data signals to each user. This scenario does not require phase synchronization between BSs and therefore has the same backhaul requirements as conventional Massive MIMO systems, where each user is preassigned to only one BS. The scenario where all BSs are phase synchronized to send the same data is also included for comparison. We solve a total transmit power minimization problem in order to observe how much power Massive MIMO BSs consume to provide the requested quality of service (QoS) of each user. A max-min fairness optimization is also solved to provide every user with the same maximum QoS regardless of the propagation conditions.
The second part of the thesis considers a joint pilot design and uplink (UL) power control problem in multi-cell Massive MIMO. The main motivation for this work is that the pilot assignment and pilot power allocation is momentous in Massive MIMO since the BSs are supposed to construct linear detection and precoding vectors from the channel estimates. Pilot contamination between pilot-sharing users leads to more interference during data transmission. The pilot design is more difficult if the pilot signals are reused frequently in space, as in Massive MIMO, which leads to greater pilot contamination effects. Related works have only studied either the pilot assignment or the pilot power control, but not the joint optimization. Furthermore, the pilot assignment is usually formulated as a combinatorial problem leading to prohibitive computational complexity. Therefore, in the second part of this thesis, a new pilot design is proposed to overcome such challenges by treating the pilot signals as continuous optimization variables. We use those pilot signals to solve different max-min fairness optimization problems with either ideal hardware or hardware impairments.
The third part of this thesis studies a two-layer decoding method that mitigates inter-cell interference in multi-cell Massive MIMO systems. In layer one, each BS estimates the channels to intra-cell users and uses the estimates for local decoding within the cell. This is followed by a second decoding layer where the BSs cooperate to mitigate inter-cell interference. An UL achievable SE expression is computed for arbitrary two-layer decoding schemes, while a closed form expression is obtained for correlated Rayleigh fading channels, maximum-ratio combining (MRC), and largescale fading decoding (LSFD) in the second layer. We formulate a sum SE maximization problem with both the data power and LSFD vectors as optimization variables. Since the problem is non-convex, we develop an algorithm based on the weighted minimum mean square error (MMSE) approach to obtain a stationary point with low computational complexity.
Motivated by recent successes of deep learning in predicting the solution to an optimization problem with low runtime, the fourth part of this thesis investigates the use of deep learning for power control optimization in Massive MIMO. We formulate the joint data and pilot power optimization for maximum sum SE in multi-cell Massive MIMO systems, which is a non-convex problem. We propose a new optimization algorithm, inspired by the weighted MMSE approach, to obtain a stationary point in polynomial time. We then use this algorithm together with deep learning to train a convolutional neural network to perform the joint data and pilot power control in sub-millisecond runtime. The solution is suitable for online optimization.
Finally, the fifth part of this thesis considers a large-scale distributed antenna system that serves the users by coherent joint transmission called Cell-free Massive MIMO. For a given user set, only a subset of the access points (APs) is likely needed to satisfy the users' performance demands. To find a flexible and energy-efficient implementation, we minimize the total power consumption at the APs in the DL, considering both the hardware consumed and transmit powers, where APs can be turned off to reduce the former part. Even though this is a nonconvex optimization problem, a globally optimal solution is obtained by solving a mixed-integer second-order cone program (SOCP). We also propose low-complexity algorithms that exploit group-sparsity or received power strength in the problem formulation.
@phdthesis{diva2:1376297,
author = {Van Chien, Trinh},
title = {{Spatial Resource Allocation in Massive MIMO Communications:
From Cellular to Cell-Free}},
school = {Linköping University},
type = {{Linköping Studies in Science and Technology. Dissertations No. 2036}},
year = {2020},
address = {Sweden},
}
Thermal cameras have historically been of interest mainly for military applications. Increasing image quality and resolution combined with decreasing camera price and size during recent years have, however, opened up new application areas. They are now widely used for civilian applications, e.g., within industry, to search for missing persons, in automotive safety, as well as for medical applications. Thermal cameras are useful as soon as there exists a measurable temperature difference. Compared to cameras operating in the visual spectrum, they are advantageous due to their ability to see in total darkness, robustness to illumination variations, and less intrusion on privacy.
This thesis addresses the problem of automatic image analysis in thermal infrared images with a focus on machine learning methods. The main purpose of this thesis is to study the variations of processing required due to the thermal infrared data modality. In particular, three different problems are addressed: visual object tracking, anomaly detection, and modality transfer. All these are research areas that have been and currently are subject to extensive research. Furthermore, they are all highly relevant for a number of different real-world applications.
The first addressed problem is visual object tracking, a problem for which no prior information other than the initial location of the object is given. The main contribution concerns benchmarking of short-term single-object (STSO) visual object tracking methods in thermal infrared images. The proposed dataset, LTIR (Linköping Thermal Infrared), was integrated in the VOT-TIR2015 challenge, introducing the first ever organized challenge on STSO tracking in thermal infrared video. Another contribution also related to benchmarking is a novel, recursive, method for semi-automatic annotation of multi-modal video sequences. Based on only a few initial annotations, a video object segmentation (VOS) method proposes segmentations for all remaining frames and difficult parts in need for additional manual annotation are automatically detected. The third contribution to the problem of visual object tracking is a template tracking method based on a non-parametric probability density model of the object's thermal radiation using channel representations.
The second addressed problem is anomaly detection, i.e., detection of rare objects or events. The main contribution is a method for truly unsupervised anomaly detection based on Generative Adversarial Networks (GANs). The method employs joint training of the generator and an observation to latent space encoder, enabling stratification of the latent space and, thus, also separation of normal and anomalous samples. The second contribution is the previously unaddressed problem of obstacle detection in front of moving trains using a train-mounted thermal camera. Adaptive correlation filters are updated continuously and missed detections of background are treated as detections of anomalies, or obstacles. The third contribution to the problem of anomaly detection is a method for characterization and classification of automatically detected district heat leakages for the purpose of false alarm reduction.
Finally, the thesis addresses the problem of modality transfer between thermal infrared and visual spectrum images, a previously unaddressed problem. The contribution is a method based on Convolutional Neural Networks (CNNs), enabling perceptually realistic transformations of thermal infrared to visual images. By careful design of the loss function the method becomes robust to image pair misalignments. The method exploits the lower acuity for color differences than for luminance possessed by the human visual system, separating the loss into a luminance and a chrominance part.
@phdthesis{diva2:1365154,
author = {Berg, Amanda},
title = {{Learning to Analyze what is Beyond the Visible Spectrum}},
school = {Linköping University},
type = {{Linköping Studies in Science and Technology. Dissertations No. 2024}},
year = {2019},
address = {Sweden},
}
Last updated: 2020-10-01