| 1. |
- Dadhich, Siddharth, 1987-
(författare)
-
Automation of Wheel-Loaders
- 2018
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Automation and tele-remote operation of mobile earth moving machines is desired for safety and productivity reasons. With tele-operation and automation, operators can avoid harsh ergonomic conditions and hazardous environments with poor air quality, and the productivity can in principle be improved by saving the time required to commute to and from work areas. Tele-remote operation of a wheel-loader is investigated and compared with manual operation, and it is found that the constrained perception of the machine is a challenging problem with remote operations. Real-time video transmission over wireless is difficult, but presents a way towards improving the remote operator’s quality of experience. To avoid glitches in the real-time video, arising from variable wireless conditions, the use of SCReAM (Self-Clocked Rate Adaptation for Multimedia) protocol is proposed. Experiments with a small scale robot over LTE show the usefulness of SCReAM for time-critical remote control applications. Automation of the bucket-filling step in the loading cycle of a wheel-loader has been an open problem, despite three decades of research. To address the bucket-filling problem, imitation learning has been applied using expert operator data, experiments are performed with a 20-tonne Volvo L180H wheel-loader and an automatic bucket-filling solution is proposed, developed and demonstrated in field-tests. The conducted experiments are in the realm of small data (100 bucket-filling examples), shallow time-delayed neural-network (TDNN), and a wheel-loader interacting with a non-stationary pile-environment. The total delay length of the TDNN model is found to be an important hyperparameter, and the trained and tuned model comes close to the performance of an expert operator with slightly longer bucket-filling time. The proposed imitation learning trained on medium coarse gravel succeeds in filling buckets in a gravel cobble pile. However, a general solution for automatic bucket-filling needs to be adaptive to possible changes in operating conditions. To adapt an initial imitation model for unseen operating conditions, a reinforcement learning approach is proposed and evaluated. A deterministic actor-critic algorithm is used to update actor (control policy) and critic (policy evaluation) networks. The experiments show that by use of a carefully chosen reward signal the models learns to improve and maximizes bucket weights in a gravel-cobble pile with only 40 bucket-filling trials. This shows that an imitation learning based bucket-filling solution equipped with a reinforcement learning agent is well suited for the continually changing operating conditions found in the construction industry. The results presented in this thesis are a demonstration of the use of artificial intelligence and machine learning methods for the operation of construction equipment. Wheel-loader OEMs can use these results to develop an autonomous bucket-filling function that can be used in manual, tele-remote or fully autonomous operations.
|
|
| 2. |
- Hartman, Andreas
(författare)
-
Electromagnetic Modeling with Complex Dielectrics : A Partial Element Equivalent Circuit Approach
- 2019
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Wireless communication systems have become an integral part of many complex systems in diverse areas of society, for the exchange of data in business and industrial settings. With the advent of Internet of Things (IoT) and wireless sensor network architectures, the tighter demands on interoperability between different devices are putting heavy requirements their ability to exchange data wirelessly among them reliably. However, many environments pose a challenging setting for a wireless communication system to operate within. Consequently, electromagnetic modeling could be used as a crucial part of the analysis and design of a wireless communication system in these environments.In this thesis, means for the electromagnetic modeling of complex materials are considered. Specifically, the incorporation of dielectrics that exhibit loss, dispersion, and anisotropic properties into electromagnetic codes is addressed. The work has been executed within the partial element equivalent circuit (PEEC) method framework.First, a PEEC implementation that incorporates dispersive and lossy dielectrics, represented by equivalent circuit models explicitly included in the PEEC equations, is developed. This provides a descriptor system form of the PEEC model that includes dielectrics with permittivities that can be represented as finite sums of Debye and Lorentz permittivity models and can be integrated by any time integration scheme of choice. Additionally, the description admits the application of model-order reduction techniques, reducing the model complexity of a large-scale PEEC model that consists of frequency-dispersive dielectrics.Next, the incorporation of anisotropic dielectrics in PEEC simulations is considered. A PEEC cell for anisotropic dielectrics, with a general permittivity tensor, is derived. It turns out to be an extension of the standard dielectric PEEC cell for an isotropic dielectric by adding a voltage-dependent current source in parallel with the excess capacitance of the dielectric cell. A cross-coupling excess capacitance concept that defines the dependent current source for the anisotropic PEEC cell is defined and given for orthogonal PEEC meshes. As a result, the PEEC cell for an anisotropic dielectric is possible to extend to handle lossy and dispersive anisotropic dielectrics straightforwardly. The developed PEEC model has been applied to model a patch antenna mounted on an anisotropic substrate. The simulation results are in agreement with other simulation technique results. Consequently, the anisotropic model permits electromagnetic modeling of structures and devices that consist of a broader class of materials.The modeling of dielectrics in different ambient temperature conditions is also considered for the PEEC analysis of its impact on antennas. Dielectrics with temperature dependent permittivity have been modeled with PEEC by standard approaches found in the literature. This has proved useful for frequency-domain simulations in PEEC. The utility has been demonstrated by investigating the impact due to temperature-dependent dielectrics on printed antennas. These types of investigations could provide valuable in-formation in the design of printed antennas in harsh environments.Finally, the problem of designing magneto-dielectric materials that intrinsically provide distortionless propagation for TEM mode signals is investigated. The frequency dependent permittivity and permeability of a slab are related to the per-unit length (p.u.l.) parameters of a two-conductor transmission line. The p.u.l. parameters are specified to approximate the Heaviside condition in a specified and finite frequency interval, while simultaneously enforcing that the corresponding permittivity and permeability represent a passive material. Consequently, the passivity condition ensures the designed material is possible to realize in practice while the Heaviside condition secures that the material is distortionless. The design method has been employed to design a passive material that approximates the Heaviside condition in a narrow frequency interval. Verification in both time and frequency domains indicates that the designed material closely resembles a distortionless material in the specified frequency interval. These results indicate that an approximation of the Heaviside condition could be a potential aid in the design of distortionless materials for bandlimited applications. Further investigations on design method improvements, limitations on the approximation in terms of both accuracy and bandwidth, and the construction of such materials in practice could lead to new distortionless cable or material designs.
|
|
| 3. |
- Häggström, Fredrik, 1984-
(författare)
-
Robust energy management for IoT machine elements
- 2018
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Advancements in semiconductor technology have reached the point at which ambient energy in the surrounding environment can power sensors and microprocessors. This capability enables new strategies for energy management, which are necessary to continue the sensorization of our environment. With the vast amount of interconnected devices and the rate at which the number of such devices is increasing, there is a need to power resource-constrained devices through means other than disposable batteries. Harvesting ambient energy from the surroundings of the device is one solution to this challenge. It has been estimated that the global demand for bearings will reach 104.5 billion dollars in 2018 with an annual increase of 7.3%. Bearings are mechanical parts that are essential for rotating machinery and that have the potential to measure and monitor vital parts of a machine. A scenario in which bearings contain embedded electronics to monitor process and health parameters that can be analyzed on site and collected from remote locations is a crucial motivator for this thesis. The investigated technologies should be applicable in dirty and encapsulated industrial environments; therefore, vibrational and rotational kinetic energies are considered in this thesis. For each energy source, both the physics and the associated electronics are modeled and to some extent experimentally verified. Vibration harvesters are investigated and modeled in SPICE to verify performance gains using a novel circuit for nonlinear power extraction for piezoelectric materials. The simulations revealed that a weak coupling from the electrical system to the mechanical system would greatly benefit nonlinear extraction techniques. Such a weakly coupled system can be created in a bearing. Mechanical load and rotation generate cyclic strain in the bearing's raceway; the cyclic strain can be utilized by applying piezoelectric patches to the raceway to power embedded systems, and sensory information from the piezoelectric patch can also be used to monitor the bearing. Finally, trends and limits for the energy costs of computing, communication and data acquisition are investigated to determine suitable energy storage technologies to combine with the advancements in energy harvesting for machine elements such as rolling element bearings. The results indicate that high integration between the mechanical and electrical parts is desired, which, in combination with capacitive energy storage, appears to be the long-term direction for real-world implementations.
|
|
