**2. Materials and Methods**

In this section, we present the methodology used to design the user intent recognition (UIR) system. The architecture of the UIR system is illustrated in Figure 1. Our new contribution is a novel feature selection method based on multi-objective optimization (MOO), as illustrated in Figure 1 in the double-lined box (Section 2.4). In this box, the application of four multi-objective biogeography-based optimization (MOBBO) methods for gait mode classification is new, and a novel MOO-based feature selection method called gradient-based multi-objective feature selection (GMOFS) is new. The remaining parts of the UIR system are implemented based on the existing literature. The role of each subsystem is explained in more detail in the following subsections.

**Figure 1.** Architecture of user intent recognition system. The double-lined box indicates that an evolutionary algorithm is used for optimization.
