A systematic review of functions and design features of in-vehicle agents

• Interests in in-vehicle agents (IVAs) have sharply increased. • We reviewed 61 articles with two research questions about functions and design features of in-vehicle agents using the PRISMA framework. • We provide functional categories and elicit design space, variables, and features of IVAs. • We discuss how we can understand the interaction with IVAs based on human-agent interaction theories. Intelligent agents are expected to become major computing systems in the near future. In-vehicle agents (IVAs) can be widely adopted and utilized in a driving environment. Despite their potential applicability, a holistic understanding of the functional objectives and design features of IVAs remains limited, which hinders the design of better IVAs. To this end, we aimed at a better understanding of the different functions of IVAs and analyzed the design features using a systematic review approach. A total of 61 articles were finally selected for review. The selected articles were analyzed according to the research questions. As a result, we provided functional categories of IVAs and elicited design space, variables, and levels. Our work contributes to the body of knowledge on the development and design of IVAs, particularly for both practitioners and researchers to conceptualize, implement, and develop their own IVAs.