Use of parallel recognizers for robust in-car speech interaction