Efficient Augmented Reality on Low-Power Embedded Systems

In this paper we propose a development technique for low-power devices with limited computing capacity to obtain efficient, high-performance and non-CPU-invasive Augmented Reality (AR) applications. The paper will discuss how to exploit both the available hardware and software resources. Many boards on the market are equipped with CPUs with low computing power together with GPUs for 2D/3D graphics and multimedia. The paper analyses the strengths of these architectures and how to exploit them. The Operating System (O.S.) also provides features that allow greater control over the system (e.g., avoid wasting resources) and its performance. The techniques proposed are then used, as an example, in the development of an AR application for remote assistance.