The MIT voyager speech understanding system, described in some detail in [1], is an urban exploration system that interacts with the user through spoken dialogue and graphics. This paper describes some recent changes made in the Voyager system, and attempts to assess the effectiveness of these developments. Two key improvements are a tighter integration of the speech and natural language components and a pipelined hardware implementation leading to a speed-up in processing time from approximately 12 times real time to approximately 3 times real time. We have also made a number of incremental improvements in the word-pair grammar, pronunciation networks, and the back-end capabilities. Finally, we demonstrate substantial performance improvements, mainly as a consequence of tighter integration.