“…These factors lead to the advantage that most LME methods can be utilized on‐line in complete unknown environments. There are a lots of LME methods, such as the combination of searching algorithm and the APF method (Sharifi and Vinke, 1993; Barraquand et al , 1992; Barraquand and Latombe, 1991), virtual obstacle (VO) method (Park and Lee, 2003; Li et al , 2000), sub‐goal method (Bell and Weir, 2004; Weir et al , 2006), wall‐following method (Borenstein and Koren, 1989; Yun and Tan, 1997; Park et al , 2005; Mabrouk and McInnes, 2008), etc. Nevertheless, since heuristic strategies only based on local environment information are usually applied, many LME methods cannot guarantee to converge in general.…”