UC Davis Institute of Transportation Studies

While the existing work on technology innovation is abundant, the innovation process largely remains a “black box,” shrouded in mystery. Energy models that incorporate innovation concepts, such as experience curves, fail to consider the fundamental processes that drive innovation. This dissertation research establishes a set of methodological approaches to better break in to this innovation black box, aiding in the quantification of the more qualitative approaches to innovation. These methods are applied to better examine low-carbon technology innovation in transportation. Specifically, this dissertation looks at biofuel innovation and the more recent diffusion of electric vehicles. Patent trends, one traditional approach for quantifying innovations, are used to provide a point of comparison for the novel methodologies employed. This research shows that the innovation narrative and conclusions that can be drawn from patent data are largely dependent on how patents are classified. Employing statistical models in conjunction with computational linguistics and machine-learning algorithms, it is possible to classify large bodies of text. This methodology is applied to a large selection of patents to better classify biofuel technologies. Additionally, this method is applied to a large repository of textual media, such as newspaper articles and trade journals, to select for specific technologies, and to classify articles by the type of information they convey. This Technology Innovation System (TIS) database is believed to adequately proxy the flow of information over time, due to the large number of documents collected. The innovation trends captured in the TIS database align well with the biofuel narrative established in literature. There is also good alignment between patent data classified through this methodology and the TIS database. Through use of the TIS database in conjunction with deployment data and policy data, this dissertation demonstrates several applications for assessing technology innovation. Results can be used to provide suggestions, supported by the data, which may foster improved innovation outcomes for low-carbon transportation technologies