Artificial intelligence (AI) and unmanned driving technologies have become the hotspots in the development of science and technology. The emergence of unmanned vehicles have provided a convenient way to travel and transport goods. In this paper, a small-sized intelligent indoor unmanned vehicle for buildings logistics is designed and implemented. The chassis of four-wheel differential steering vehicle is used as the basic of driving system, and the ROS operating platform is built as the core control system of the intelligent vehicle; For the environment perception system, the RGB camera and 20-line LiDAR are used as environment sensors and sets of environment data are fused and optimized through fuzzy fusion algorithm and DNN neural network; Meanwhile，the intelligent unmanned logistics vehicle obtains the information of position, attitude and speed by 20-line LiDAR and inertial navigation; Furthermore, the SLAM (simultaneous localization and mapping) algorithm is used to realize indoor path planning, collision avoidance and indoor navigation. In the real tested scenario, the intelligent unmanned logistics vehicle can achieve wireless communication with elevator controller through internal WiFi module, and transport goods across any floors; Finally, In the real tested scenario, the results have shown that the intelligent vehicle not only can complete indoor map modeling and path planning in a closed environment，but also it can realize dynamic detection of obstacles and active collision avoidance. During the actual cross-floor delivery process, the intelligent unmanned logistics vehicle can accurately and efficiently complete the complex indoor delivery tasks.

CICTP