No description available.

Natural language (NL) is a common medium humans use to express ideas and communicate with others, while programming languages (PL) are the ``language'' humans use to communicate with machines. As NL and PL were designed for different purposes, a considerable difference exists in the structure and capabilities. Programming using PL can take novices months to learn. Meanwhile, users are already familiar with NL. Therefore, natural language programming (NLPr) holds excellent potential by giving non-experts the ability to ``program'' with the language they already know and a Low-Code/No-Code development experience. However, many challenges with developing NLPr systems are yet to be addressed, namely how to disambiguate NL semantics, validate inputs and provide helpful feedback, and generate the executable programs based on semantic meanings effectively. This dissertation addresses these issues by proposing a Controlled Object-Oriented Language (COOL) model to disambiguate and analyze the English inputs' semantic meanings and implement a LEGO robot NLPr platform. Two main approaches that connect the current research in general-purpose NLP to NLPr are taken: (1) A domain-specific lexicon and function library serve as the syntax and semantic space. Even though NL can be complex and expressive, functions for the specific robot domain can be fulfilled with libraries built of a finite set of objects and functions. (2) An error-reporting and feedback mechanism detects erroneous sentences, explains possible reasons, and provides debugging and rewriting suggestions. The error-reporting and feedback systems are developed with a hybrid approach that combines rule-based methods such as FSM and dependency-based structural analysis with the data-based multi-label classification (MLC) method. Experiment results and user studies show that, with the proposed model and approaches reducing the ambiguity within the target domain, the NLPr system can process a relatively expressive controlled NL for robot motion control and generate executable codes based on the English input. When the system is confronted with erroneous sentences, it produces error messages, suggestions, and example sentences for users. NL's structural and semantic information can be transformed into the intermediate representations used for program synthesis with the language model and system proposed to resolve the situation where the considerable amount of data needed for a data-based model is unavailable. / Doctor of Philosophy / Natural language (NL) is one of the most common mediums humans use daily to express and explain ideas and communicate with each other. In contrast, programming languages (PL) are the ``language'' humans use to communicate with machines. Because of the difference in the purpose, media, and audience, there is a considerable difference in their structure and capabilities. NL is more expressive and natural and sometimes can be rather complex, while PL is primarily short, straightforward, and not as expressive as NL. The need for programming has increased in recent years. However, the learning curve of programming languages can easily be months or more for novice users to learn. At the same time, all potential users are familiar with at least one NL. As such, natural language programming (NLPr), a technology that enables people to program with NL, holds excellent potential since it gives non-experts the ability to ``program'' with the language they already know and a Low-Code or even No-Code development experience. However, despite recent research into NLPr, many challenges with developing NLPr systems are yet to be addressed, namely how to disambiguate natural language semantics, how to validate inputs and provide helpful feedback with a limited amount of data, and how to effectively generate the executable programs based on the semantic meanings. This dissertation addresses these issues by proposing a Controlled Object-Oriented Language (COOL) model to disambiguate and analyze the English inputs' semantic meanings and implement a LEGO robot NLPr platform. Two main approaches that connect the current research in general-purpose NLP techniques to NLPr are taken: (1) The first is developing a domain-specific lexicon and function library with the designed COOL model to serve as the syntax and semantic space. Even though natural language can be extremely complex and expressive, the functions for the specific robot domain can be fulfilled with libraries built of a finite set of objects and functions. (2) An error-reporting and feedback mechanism detects erroneous sentences, explains possible reasons, and provides debugging and rewriting suggestions. The error-reporting and feedback systems are developed with a hybrid approach that combines rule-based methods such as FSM and dependency-based structural analysis with the data-based multi-label classification (MLC) method. Experiment results and user studies show that, with the proposed language model and approaches reducing the ambiguity within the target domain, the designed NLPr system can process a relatively expressive controlled natural language designed for robot motion control and generate executable codes based on the semantic information extracted. When the NLPr system is confronted with erroneous sentences, it produces detailed error messages and provides suggestions and sample sentences for possible fixes to users. NL's structural and semantic information can be transformed into the intermediate representations used for program synthesis with the simple language model and system proposed to resolve the situation where the considerable amount of data needed for a data-based model is unavailable.

This thesis investigates the application of index based, region segmentation, and deep learning methods to the sensor fusion of omnidirectional (O-D) Infrared (IR) sensors, Kinnect sensors, and O-D vision sensors to increase the level of intelligent perception for unmanned robotic platforms. The goals of this work is first to provide a more robust calibration approach and improve the calibration of low resolution and noisy IR O-D cameras. Then our goal was to explore the best approach to sensor fusion for vegetation detection. We looked at index based, region segmentation, and deep learning methods and compared them with a goal of significant reduction in false positives while maintaining reasonable vegetation detection. The results are as follows: Direct Spherical Calibration of the IR camera provided a more consistent and robust calibration board capture and resulted in the best overall calibration results with sub-pixel accuracy The best approach for sensor fusion for vegetation detection was the deep learning approach, the three methods are detailed in the following chapters with the results summarized here. Modified Normalized Difference Vegetation Index approach achieved 86.74% recognition and 32.5% false positive, with peaks to 80% Thermal Region Fusion (TRF) achieved a lower recognition rate at 75.16% but reduced false positives to 11.75% (a 64% reduction) Our Deep Learning Fusion Network (DeepFuseNet) results demonstrated that deep learning approach showed the best results with a significant (92%) reduction in false positives when compared to our modified normalized difference vegetation index approach. The recognition was 95.6% with 2% false positive. Current approaches are primarily focused on O-D color vision for localization, mapping, and tracking and do not adequately address the application of these sensors to vegetation detection. We will demonstrate the contradiction between current approaches and our deep sensor fusion (DeepFuseNet) for vegetation detection. The combination of O-D IR and O-D color vision coupled with deep learning for the extraction of vegetation material type, has great potential for robot perception. This thesis will look at two architectures: 1) the application of Autoencoders Feature Extractors feeding a deep Convolution Neural Network (CNN) fusion network (DeepFuseNet), and 2) Bottleneck CNN feature extractors feeding a deep CNN fusion network (DeepFuseNet) for the fusion of O-D IR and O-D visual sensors. We show that the vegetation recognition rate and the number of false detects inherent in the classical indices based spectral decomposition are greatly improved using our DeepFuseNet architecture. We first investigate the calibration of omnidirectional infrared (IR) camera for intelligent perception applications. The low resolution omnidirectional (O-D) IR image edge boundaries are not as sharp as with color vision cameras, and as a result, the standard calibration methods were harder to use and less accurate with the low definition of the omnidirectional IR camera. In order to more fully address omnidirectional IR camera calibration, we propose a new calibration grid center coordinates control point discovery methodology and a Direct Spherical Calibration (DSC) approach for a more robust and accurate method of calibration. DSC addresses the limitations of the existing methods by using the spherical coordinates of the centroid of the calibration board to directly triangulate the location of the camera center and iteratively solve for the camera parameters. We compare DSC to three Baseline visual calibration methodologies and augment them with additional output of the spherical results for comparison. We also look at the optimum number of calibration boards using an evolutionary algorithm and Pareto optimization to find the best method and combination of accuracy, methodology and number of calibration boards. The benefits of DSC are more efficient calibration board geometry selection, and better accuracy than the three Baseline visual calibration methodologies. In the context of vegetation detection, the fusion of omnidirectional (O-D) Infrared (IR) and color vision sensors may increase the level of vegetation perception for unmanned robotic platforms. A literature search found no significant research in our area of interest. The fusion of O-D IR and O-D color vision sensors for the extraction of feature material type has not been adequately addressed. We will look at augmenting indices based spectral decomposition with IR region based spectral decomposition to address the number of false detects inherent in indices based spectral decomposition alone. Our work shows that the fusion of the Normalized Difference Vegetation Index (NDVI) from the O-D color camera fused with the IR thresholded signature region associated with the vegetation region, minimizes the number of false detects seen with NDVI alone. The contribution of this work is the demonstration of two new techniques, Thresholded Region Fusion (TRF) technique for the fusion of O-D IR and O-D Color. We also look at the Kinect vision sensor fused with the O-D IR camera. Our experimental validation demonstrates a 64% reduction in false detects in our method compared to classical indices based detection. We finally compare our DeepFuseNet results with our previous work with Normalized Difference Vegetation index (NDVI) and IR region based spectral fusion. This current work shows that the fusion of the O-D IR and O-D visual streams utilizing our DeepFuseNet deep learning approach out performs the previous NVDI fused with far infrared region segmentation. Our experimental validation demonstrates an 92% reduction in false detects in our method compared to classical indices based detection. This work contributes a new technique for the fusion of O-D vision and O-D IR sensors using two deep CNN feature extractors feeding into a fully connected CNN Network (DeepFuseNet).