“There’s no one in the whole haulage industry who says, damn, that was well driven” - An Interaction Design Study of Eco-Driving Support for Professional Truck Drivers

Abstract

This thesis explores how eco-driving support for professional truck drivers can be designed from an interaction design perspective. Existing eco-driving systems in the commercial vehicle domain are often retrospective, score-based, visually oriented, and weakly contextualised, which may limit their usefulness in the safety-critical, demanding, and organisationally complex context of truck driving. At the same time, recent advances in sensing, connectivity, voice interaction, and large language models (LLMs) have created new possibilities for more adaptive driver coaching. However, research on such emerging concepts in the truck eco-driving domain remains fragmented. Conducted in collaboration with Volvo Group Trucks Technology &Industrial Division (TTI), the thesis emerged from an ongoing exploration of these possibilities, but approached them through a user-centred interaction design process in order to ground them in truck drivers’ needs and real use conditions. Against this background, the thesis addressed the following exploratory research question: What can an interaction design lens reveal about the challenges and opportunities of designing eco-driving support for professional truck drivers? The design process was inspired by the first three phases of the Double Diamond: Discover, Define, and Develop. Methods included literature review, industry benchmarking, semi-structured interviews with truck drivers and fleet managers, reflexive thematic analysis, ideation, prototyping, and prototype evaluation. The resulting mid-fidelity prototype applied generative LLM capabilities across both on-board and off-board components, combining a user-initiated, voice-based eco-driving coach with a manager-facing view for contextualised follow-up and recognition. The findings show that eco-driving support should be understood as a socio-technical and organisational design problem rather than only an interface problem. Four main results were identified: organisational conditions shape how eco-driving support is experienced; one-size-fits-all coaching is inadequate in diverse driving contexts; eco-driving feedback must balance tone, explainability, timing, and flow control; and user acceptance functions as the final filter for system success. The study also highlights several tensions between context-aware adaptation and privacy, explainability and limited attention, and feedback timing and driver control. Rather than validating AI or voice interaction as definitive solutions, the thesis contributes user-centred design knowledge about the conditions, tensions, and opportunities that future eco-driving support could address in professional truck driving.