A profound transformation of the global energy system is underway and the next few decades will see a fifty fold expansion in photovoltaic (PV) solar generation. The vast scale of PV deployment necessarily means it cannot be considered in isolation, separate from the ecological, economic or social systems in which it is deployed. At this juncture exists the opportunity to integrate the required PV generation in an environmentally and socially responsible way. Agrivoltaics, the co-location of PV and agriculture where the PV is adapted to needs of the agricultural production, offers particular opportunity. If done thoughtfully, agrivoltaics can improve the sustainability of some of our most essential resources: energy, food, water and land. This thesis reviews agrivoltaic requirements, then considers how the fundamental generation unit of solar energy, the PV module, could best be adapted for use with horticultural production and maximise benefits to both. A novel semi-transparent agrivoltaic specific PV module is proposed, herein termed a low concertation agrivoltaic panel (LAP). Based on a matrix of crystalline silicon (c-Si) cells along with embedded optics to provide appropriate spectral selectivity, the module allows photosynthetically active radiation (PAR) to pass while redirecting near infrared (NIR) radiation to the cells for electrical generation. To achieve spectral selectivity while enabling reflected irradiance to reach the PV cells a thin film interference dichroic is applied to planar concentrator geometry. Semi- transparency can be achieved with other PV materials (e.g. a-Si, OPV, DSSC), or architectures employed to achieve specular selectivity (e.g. luminescent solar concentrators). A review of these for agrivoltaic applications is undertaken, none show performance close to the benchmark of a module partially populated with opaque c-Si cells, and in many cases stability and lifetime challenges still need to be overcome. To date a commercially relevant, spectrally selective PV module has yet to be demonstrated for use in agrivoltaic applications. To enable design refinement of the LAP concept and simulation of its performance across representative sites, orientations and solar resource conditions, a detailed optical model based on ray tracing is developed. Modelled results are used to estimate the electrical generation advantage of the LAP concept over a typical partially populated c-Si module with equivalent cell coverage area. Incorporating LAP embedded optics to a module with 50% cell coverage could provide an increase of 25% in short circuit current on an annual average basis. If the desired cell coverage is reduced to 38%, an additional 40% in short circuit current can be made available. In both cases over 90% of the original PAR transmission can be maintained and crops shielded from approximately 80% of NIR radiation, reducing crop heating and consequently water consumption. Concept designs optimised via the model are used as a starting point to detail full scale designs for a suite of LAP modules with different PAR transparencies. Panel sizing, cell layout, electrical connection and component definition were performed to demonstrate the feasibility of producing a LAP module that follows contemporary PV module design and, in the most part, uses standard PV module componentry. Small scale LAP prototypes, of aperture 214 x 168 mm and containing three 30mm wide cell strips at a cell coverage of 50%, were manufactured. Testing of these under solar simulation and on sun in clear and overcast conditions confirmed that the optical model is suitably representative. On sun results for a prototype employing a commercially available dichroic film showed a short circuit current increase of 30% at normal incidence under clear skies, while maintaining 95% of the original PAR transmission. In overcast conditions an improvement of 25% in short circuit current is observed while maintaining 74% of PAR transmission. Finally, possible methods for commercial production of the embedded dichroic and next steps in the development process are discussed. As the proposed LAP concept leverages incumbent c-Si cell technology, components and manufacturing in a contemporary monolithic PV module format, it has the potential to be quickly commercialised at low incremental cost. Significant electrical generation benefits with minimal loss in PAR transmission demonstrated in this thesis provide justification for further effort to develop the concept and address the emerging agrivoltaics market.