We present our work on 3D path planning for a mobile robot in unknown and cluttered environments. The hardware platform is composed of a tracked rover with four flippers which enable the robot to crawl over and under obstacles, climb stairs and traverse various terrain including sand, gravel, muck, snow, and ice. This research is motivated by use-cases for this remotely operated mobile robot in applications ranging from disaster response to sample acquisition to environment monitoring. Even though there is a trained human operator in the loop controlling the robot, the uncertain and densely structured environments pose operational challenges. Our work extends a path planning method known as “driving with tentacles” to 3D navigation in a cluttered environment with complicated structures including coils, risers and interlaced pipes. We will present results from various mission scenarios including obstacle avoidance, stair climbing, and crawling over obstacles.