Search

Submission note: A thesis submitted in total fulfillment for the degree of Doctor of Philosophy to the Department of Electronic Engineering, Faculty of Science, Technology and Engineering, La Trobe University.This thesis presents a novel sensorless position technique for a Switched Reluctance Motor (SRM). The widespread use of SRMs has been relatively limited when compared to other popular motor technologies. Renewed interest has occurred in SRMs due to low cost, environmentally friendly materials and its wide torque operating region. One of the disadvantages affecting the wider use of an SRM, is the requirement for a rotor position sensor. This increases the cost and reduces the durability of the system. As a result, much research has been dedicated to making the motor `sensorless'. The proposed sensorless technique seeks to improve the operating range of currently available methods, with the advantage that there is no need for an offline calibration technique. To achieve sensorless operation, the proposed technique evaluates the inductance waveform on a per phase basis, to find the minimum and maximum inductance positions. These locations correspond to the unaligned and aligned rotor positions respectively. A sensorless rotor position is created via a novel digital phase locked loop (DPLL), which generates intermediate positions between these detected positions. To enhance sensorless tracking and accuracy a multiple model adaptive estimator Kalman filter is introduced. This improves the position estimates of the un-aligned and aligned rotor positions and improves the sensorless tracking output of the DPLL. Lastly, to ensure sensorless operation from zero speed to normal sensorless operation, a novel hesitation-free starting technique is presented. The performance of the proposed sensorless technique is evaluated using simulations and real-world testing utilising a three-phase test SRM. Novel methods have been developed to remove the effects of mutual coupling, flux-linkage term errors and magnetic hysteresis. These techniques ensure that accurate sensorless operation is achievable over a wide range of speeds and loads, without the need for an offline calibration technique. This results in improved durability and reduced cost of the SRM.

In recent years, the capabilities of legged locomotion controllers have been significantly advanced enabling them to traverse basic types of uneven terrain without visual perception. However, safely and autonomously traversing longer distances over difficult uneven terrain requires appropriate motion planning using online collected environmental knowledge. In this paper, we present such a novel methodology for generic closed-loop preceding horizon footstep planning that enables legged robots equipped with capable locomotion controllers to autonomously traverse previously unknown terrain while continuously walking long distances. Hereby, our approach addresses the challenge of online terrain perception and soft real-time footstep planning. The proposed new formulation of the search-based planning problem makes no specific assumptions about the robot kinematics (e.g. number of legs) or the used locomotion control schemes. Therefore, it can be applied to a broad range of different types of legged robots. Unlike current methods, the proposed new framework can optionally consider the floating base as part of the state-space. It is possible to configure the complexity of the planner online, from efficiently solving tasks in flat terrain to using non-contiguous contacts in highly challenging terrain. Finally, the presented methodology is successfully applied and evaluated in virtual and real experiments on state of the art bipedal, quadrupedal, and a novel eight-legged robot.

Among a large variety of available mobile ground robotic designs, legged robots are potentially the most versatile systems with respect to the traversal of uneven ground. However, they also have to deal with the most complex trajectory planning and control problems, when compared, e.g., to ground robots locomoting on wheels or tracks. The ability of legged robots to generate continuous motion over larger distances of uneven terrain by using discrete foot displacements leads to especially capable locomotion abilities in challenging environments. This potentially allows them to cross a large variety of non-contiguous surfaces (e.g. stepping stones) and to climb over obstacles (e.g. debris). In recent years, interest in humanoid and four-legged robots has significantly increased due to the attention and efforts initiated through the DARPA Robotics Challenge (2013-2015) leading to significant progress in legged robot hardware and motion controller performance. However, to fully benefit from the potential of these unique locomotion capabilities, advances in autonomous functionality is required as the high mechatronic complexity of such robots makes teleoperation at the joint-level by humans highly infeasible. The ability of supervision by a remote human operator on a high level of abstraction is needed to traverse previously unknown terrains with a legged robot for the first time. This thesis addresses several such advances: Efficient perception especially of uneven, rough and irregular terrain, and suitable digital representations of them are developed as mandatory prerequisites for developing new methods for finding feasible (and posturally stable) foothold positions and their alignment in such challenging terrain. On top of these, planning and navigation of sequences of suitable footsteps is developed to traverse a terrain section. Execution of footstep sequences over long distances requires methods using the available robot specific legged locomotion controllers for repeated online updating and stitching of footstep plans to achieve continuous autonomous walking over long distances in uneven terrain. Moreover, the reusability of methods for perception, planning, and step execution for different types of legged robots is highly desirable to efficiently adapt and deploy the resulting framework to different types of robots and their specific kinematic and dynamic motion properties as well as their motion controllers. The innovations, methods and their implementation presented in this work are successfully evaluated in challenging irregular terrain with seven different types of state of the art robots: two versions of Boston Dynamics' humanoid Atlas robot, two versions of the THORMANG robot by Robotis equipped with custom sensing and computing, the humanoid WALK-MAN robot, the quadruped ANYmal robot of ANYbotics and a tracked robot with four actuated flipper wheels (Telemax Hybrid by Telerob). It is successfully demonstrated how the novel footstep planning framework extends the current state of research in the field of 3D planning by applying the approaches developed in this thesis to improve the overall performance of the planner and efficiently generate suitable step plans even in previously not reasonable traversable irregular terrain. The generalizability and modularity of the proposed framework allows the flexible application with different types of legged robots but also, as demonstrated in the evaluation, generalizes beyond the intended primary application purpose, e.g. as automatic flipper control for a tracked robot crossing rough terrain.

Endogenous cannabinoids are lipid-based ligands of the main cannabinoid receptors type 1 and 2 (CB1R and CB2R). In contrast to the well-studied effects of CB1Rs, much less is known about the physiological role of CB2Rs in the central nervous system (CNS). In fact, CB2Rs were considered as peripheral cannabinoid receptors representing the complementary cannabinoid receptor to the CB1R in the CNS. However recent pharmacological, behavioral and genetic studies have determined the presence of functional CB2Rs in the brain and their involvement in various physiological and pathological conditions. Endocannabinoids are produced in an activity-dependent manner. Next to their well-described role as retrograde modulators of synaptic transmission, endocannabinoids also mediate a cell-autonomous slow self-inhibition (SSI). Action potential-driven endocannabinoid production followed by binding to cannabinoid receptors induces a long-lasting hyperpolarization of the membrane potential, rendering the cell less excitable. Several studies described different endocannabinoid receptors and cellular mechanisms by which SSI is implemented in different cell types and brain areas: While pyramidal cells and interneurons in the somatosensory cortex were reported to mediate SSI via CB1Rs and G protein-coupled inward rectifying K+ (GIRK) channels, hippocampal principal cells were shown to mediate SSI in a CB2R mediated and input-resistance-independent manner. However, the molecular mechanism by which the long-lasting hyperpolarization is implemented was not clear. During my thesis I was part of a team that demonstrated that hippocampal SSI is mediated via activation of the Na+/bicarbonate cotransporter. In order to get further insight in the occurrence of SSI in different classes of neurons, we analyzed the presence of SSI in one type of hippocampal interneuron and showed that oriens-lacunosum moleculare (OLM) interneurons do not express SSI. Further, we investigate SSI in different neuron types of layer 2/3 in the primary somatosensory cortex and show that regular firing cells express SSI in contrast to fast-spiking interneurons. Trains of action potentials induced a long-lasting hyperpolarization that was accompanied by a change in input resistance due to GIRK channel activation. By using cannabinoid receptor-specific pharmacology as well as transgenic mice lacking either CB1Rs or CB2Rs, we demonstrate that this effect is mediated by CB2R activation. Taken together, hippocampal and cortical SSI both represent a CB2R-dependent mechanism; however the underlying mechanism by which the hyperpolarization is implemented differs between the different brain regions. By describing an additional cellular mechanism for SSI induction, these findings add further insights on the physiological role of CB2Rs and expand our knowledge about cell type-specific differential cannabinoid signaling. Moreover, these findings suggest CB2Rs as a promising target for therapeutic approaches.