There have been multiple attempts of coupling neural networks with external memory components for sequence learning problems. Such architectures have demonstrated success in algorithmic, sequence transduction, question-answering and reinforcement learning tasks. Most notable of these attempts is the Neural Turing Machine (NTM), which is an implementation of the Turing Machine with a neural network controller that interacts with a continuous memory. Although the architecture is Turing complete and hence, universally computational, it has seen limited success with complex real-world tasks.

In this thesis, I introduce an extension of the Neural Turing Machine, the Neural Harvard Machine, that implements a fully differentiable Harvard Machine framework with a feed-forward neural network controller. Unlike the NTM, it has two different memories - a read-only program memory and a read-write data memory. A sufficiently complex task is divided into smaller, simpler sub-tasks and the program memory stores parameters of pre-trained networks trained on these sub-tasks. The controller reads inputs from an input-tape, uses the data memory to store valuable signals and writes correct symbols to an output tape. The output symbols are a function of the outputs of each sub-network and the state of the data memory. Hence, the controller learns to load the weights of the appropriate program network to generate output symbols.

A wide range of experiments demonstrate that the Harvard Machine framework learns faster and performs better than the NTM and RNNs like LSTM, as the complexity of tasks increases.

In order for a robot to solve complex tasks in real world, it needs to compute discrete, high-level strategies that can be translated into continuous movement trajectories. These problems become increasingly difficult with increasing numbers of objects and domain constraints, as well as with the increasing degrees of freedom of robotic manipulator arms.

The first part of this thesis develops and investigates new methods for addressing these problems through hierarchical task and motion planning for manipulation with a focus on autonomous construction of free-standing structures using precision-cut planks. These planks can be arranged in various orientations to design complex structures; reliably and autonomously building such structures from scratch is computationally intractable due to the long planning horizon and the infinite branching factor of possible grasps and placements that the robot could make.

An abstract representation is developed for this class of problems and show how pose generators can be used to autonomously compute feasible robot motion plans for constructing a given structure. The approach was evaluated through simulation and on a real ABB YuMi robot. Results show that hierarchical algorithms for planning can effectively overcome the computational barriers to solving such problems.

The second part of this thesis proposes a deep learning-based algorithm to identify critical regions for motion planning. Further investigation is done whether these learned critical regions can be translated to learn high-level landmark actions for automated planning.

This research introduces Roblocks, a user-friendly system for learning Artificial Intelligence (AI) planning concepts using mobile manipulator robots. It uses a visual programming interface based on block-structured programming to make AI planning concepts easier to grasp for those who are new to robotics and AI planning. Users get to accomplish any desired tasks by dynamically populating puzzle shaped blocks encoding the robot’s possible actions, allowing them to carry out tasks like navigation, planning, and manipulation by connecting blocks instead of writing code. Roblocks has two levels, where in the first level users are made to re-arrange a jumbled set of actions of a plan in the correct order so that a given goal could be achieved. In the second level, they select actions of their choice but at each step only those actions pertaining to the current state are made available to them, thereby pruning down the vast number of possible actions and suggesting only the truly feasible and relevant actions. Both of these levels have a simulation where the user plan is executed. Moreover, if the user plan is invalid or fails to achieve the given goal condition then an explanation for the failure is provided in simple English language. This makes it easier for everyone (especially for non-roboticists) to understand the cause of the failure.

In this thesis, a new approach to learning-based planning is presented where critical regions of an environment with low probability measure are learned from a given set of motion plans. Critical regions are learned using convolutional neural networks (CNN) to improve sampling processes for motion planning (MP).

In addition to an identification network, a new sampling-based motion planner, Learn and Link, is introduced. This planner leverages critical regions to overcome the limitations of uniform sampling while still maintaining guarantees of correctness inherent to sampling-based algorithms. Learn and Link is evaluated against planners from the Open Motion Planning Library (OMPL) on an extensive suite of challenging navigation planning problems. This work shows that critical areas of an environment are learnable, and can be used by Learn and Link to solve MP problems with far less planning time than existing sampling-based planners.

A critical challenge in the design of AI systems that operate with humans in the loop is to be able to model the intentions and capabilities of the humans, as well as their beliefs and expectations of the AI system itself. This allows the AI system to be "human- aware" -- i.e. the human task model enables it to envisage desired roles of the human in joint action, while the human mental model allows it to anticipate how its own actions are perceived from the point of view of the human. In my research, I explore how these concepts of human-awareness manifest themselves in the scope of planning or sequential decision making with humans in the loop. To this end, I will show (1) how the AI agent can leverage the human task model to generate symbiotic behavior; and (2) how the introduction of the human mental model in the deliberative process of the AI agent allows it to generate explanations for a plan or resort to explicable plans when explanations are not desired. The latter is in addition to traditional notions of human-aware planning which typically use the human task model alone and thus enables a new suite of capabilities of a human-aware AI agent. Finally, I will explore how the AI agent can leverage emerging mixed-reality interfaces to realize effective channels of communication with the human in the loop.

Computer vision technology automatically extracts high level, meaningful information from visual data such as images or videos, and the object recognition and detection algorithms are essential in most computer vision applications. In this dissertation, we focus on developing algorithms used for real life computer vision applications, presenting innovative algorithms for object segmentation and feature extraction for objects and actions recognition in video data, and sparse feature selection algorithms for medical image analysis, as well as automated feature extraction using convolutional neural network for blood cancer grading.

To detect and classify objects in video, the objects have to be separated from the background, and then the discriminant features are extracted from the region of interest before feeding to a classifier. Effective object segmentation and feature extraction are often application specific, and posing major challenges for object detection and classification tasks. In this dissertation, we address effective object flow based ROI generation algorithm for segmenting moving objects in video data, which can be applied in surveillance and self driving vehicle areas. Optical flow can also be used as features in human action recognition algorithm, and we present using optical flow feature in pre-trained convolutional neural network to improve performance of human action recognition algorithms. Both algorithms outperform the state-of-the-arts at their time.

Medical images and videos pose unique challenges for image understanding mainly due to the fact that the tissues and cells are often irregularly shaped, colored, and textured, and hand selecting most discriminant features is often difficult, thus an automated feature selection method is desired. Sparse learning is a technique to extract the most discriminant and representative features from raw visual data. However, sparse learning with \textit{L1} regularization only takes the sparsity in feature dimension into consideration; we improve the algorithm so it selects the type of features as well; less important or noisy feature types are entirely removed from the feature set. We demonstrate this algorithm to analyze the endoscopy images to detect unhealthy abnormalities in esophagus and stomach, such as ulcer and cancer. Besides sparsity constraint, other application specific constraints and prior knowledge may also need to be incorporated in the loss function in sparse learning to obtain the desired results. We demonstrate how to incorporate similar-inhibition constraint, gaze and attention prior in sparse dictionary selection for gastroscopic video summarization that enable intelligent key frame extraction from gastroscopic video data. With recent advancement in multi-layer neural networks, the automatic end-to-end feature learning becomes feasible. Convolutional neural network mimics the mammal visual cortex and can extract most discriminant features automatically from training samples. We present using convolutinal neural network with hierarchical classifier to grade the severity of Follicular Lymphoma, a type of blood cancer, and it reaches 91\% accuracy, on par with analysis by expert pathologists.

Developing real world computer vision applications is more than just developing core vision algorithms to extract and understand information from visual data; it is also subject to many practical requirements and constraints, such as hardware and computing infrastructure, cost, robustness to lighting changes and deformation, ease of use and deployment, etc.The general processing pipeline and system architecture for the computer vision based applications share many similar design principles and architecture. We developed common processing components and a generic framework for computer vision application, and a versatile scale adaptive template matching algorithm for object detection. We demonstrate the design principle and best practices by developing and deploying a complete computer vision application in real life, building a multi-channel water level monitoring system, where the techniques and design methodology can be generalized to other real life applications. The general software engineering principles, such as modularity, abstraction, robust to requirement change, generality, etc., are all demonstrated in this research.

Research literature was reviewed to find recommended tools and technologies for operating Unmanned Aerial Systems (UAS) fleets in an urban environment. However, restrictive legislation prohibits fully autonomous flight without an operator. Existing literature covers considerations for operating UAS fleets in a controlled environment, with an emphasis on the effect different networking approaches have on the topology of the UAS network. The primary network topology used to implement UAS communications is 802.11 protocols, which can transmit telemetry and a video stream using off the shelf hardware. Other implementations use low-frequency radios for long distance communication, or higher latency 4G LTE modems to access existing network infrastructure. However, a gap remains testing different network topologies outside of a controlled environment.

With the correct permits in place, further research can explore how different UAS network topologies behave in an urban environment when implemented with off the shelf UAS hardware. In addition to testing different network topologies, this thesis covers the implementation of building a secure, scalable system using modern cloud computation tools and services capable of supporting a variable number of UAS. The system also supports the end-to-end simulation of the system considering factors such as battery life and realistic UAS kinematics. The implementation of the system leads to new findings needed to deploy UAS fleets in urban environments.

In a collaborative environment where multiple robots and human beings are expected

to collaborate to perform a task, it becomes essential for a robot to be aware of multiple

agents working in its work environment. A robot must also learn to adapt to

different agents in the workspace and conduct its interaction based on the presence

of these agents. A theoretical framework was introduced which performs interaction

learning from demonstrations in a two-agent work environment, and it is called

Interaction Primitives.

This document is an in-depth description of the new state of the art Python

Framework for Interaction Primitives between two agents in a single as well as multiple

task work environment and extension of the original framework in a work environment

with multiple agents doing a single task. The original theory of Interaction

Primitives has been extended to create a framework which will capture correlation

between more than two agents while performing a single task. The new state of the

art Python framework is an intuitive, generic, easy to install and easy to use python

library which can be applied to use the Interaction Primitives framework in a work

environment. This library was tested in simulated environments and controlled laboratory

environment. The results and benchmarks of this library are available in the

related sections of this document.

Many researchers aspire to create robotics systems that assist humans in common office tasks, especially by taking over delivery and messaging tasks. For meaningful interactions to take place, a mobile robot must be able to identify the humans it interacts with and communicate successfully with them. It must also be able to successfully navigate the office environment. While mobile robots are well suited for navigating and interacting with elements inside a deterministic office environment, attempting to interact with human beings in an office environment remains a challenge due to the limits on the amount of cost-efficient compute power onboard the robot. In this work, I propose the use of remote cloud services to offload intensive interaction tasks. I detail the interactions required in an office environment and discuss the challenges faced when implementing a human-robot interaction platform in a stochastic office environment. I also experiment with cloud services for facial recognition, speech recognition, and environment navigation and discuss my results. As part of my thesis, I have implemented a human-robot interaction system utilizing cloud APIs into a mobile robot, enabling it to navigate the office environment, identify humans within the environment, and communicate with these humans.

Several physical systems exist in the real world that involve continuous as well as discrete changes. These range from natural dynamic systems like the system of a bouncing ball to robotic dynamic systems such as planning the motion of a robot across obstacles. The key aspects of effectively describing such dynamic systems is to be able to plan and verify the evolution of the continuous components of the system while simultaneously maintaining critical constraints. Developing a framework that can effectively represent and find solutions to such physical systems prove to be highly advantageous. Both hybrid automata and action languages are formal models for describing the evolution of dynamic systems. The action language C+ is a rich and expressive language framework to formalize physical systems, but can be used only with physical systems in the discrete domain and is limited in its support of continuous domain components of such systems. Hybrid Automata is a well established formalism used to represent such complex physical systems at a theoretical level, however it is not expressive enough to capture the complex relations between the components of the system the way C+ does.

This thesis will focus on establishing a formal relationship between these two formalisms by showing how to succinctly represent Hybrid Automata in an action language which in turn is defined as a high-level notation for answer set programming modulo theories (ASPMT) --- an extension of answer set programs in the first-order level. Furthermore, this encoding framework is shown to be more effective and expressive than Hybrid Automata by highlighting its ability in allowing states of a hybrid transition system to be defined by complex relations among components that would otherwise be abstracted away in Hybrid Automata. The framework is further realized in the implementation of the system CPLUS2ASPMT, which takes advantage of state of the art ODE(Ordinary Differential Equations) based SMT solver dReal to provide support for ODE based evolution of continuous components of a dynamic system.