Unmanned logistics vehicle based on depth learning

Abstract

The invention relates to an unmanned logistics vehicle based on depth learning. The unmanned logistics vehicle comprises a logistics vehicle body, an ultrasonic obstacle avoidance module, a binocular stereo vision obstacle avoidance module, a motor driving module, an embedded system, a power supply module, and a visual navigation processing system. The binocular stereo vision obstacle avoidance module is used for detecting a distant obstacle in a road scene. The ultrasonic obstacle avoidance module is used for detecting a near distance obstacle, and distance information of the obstacles obtained by the two modules are called obstacle avoidance information. According to the visual navigation processing system, a depth learning model trained by the sample set is used to process collected road image data, and control command information is outputted. Finally, a decision model integrates control instruction information and the obstacle avoidance information to control the motor driving module so as to realize the unmanned driving function of a logistics vehicle. According to the unmanned logistics vehicle, the installation of auxiliary equipment is not needed, the depth learning model can sense and understand a road surrounding environment through a learning sample set, and the unmanned driving function of the logistics vehicle is realized.

Description

technical field [0001] The invention belongs to the field of unmanned driving technology, and relates to an unmanned logistics vehicle based on deep learning, which is suitable for public places such as large parks, warehouses, stations, airports, and docks. Background technique [0002] With the rapid development of the logistics industry, especially the increasing transportation volume of warehousing freight, express delivery and takeaway, it has brought great development potential and huge market space for unmanned logistics vehicles. [0003] However, at present, most of the warehousing and freight transportation use AGV (unmanned guided vehicle) with electromagnetic guidance, tape guidance and laser navigation. Express delivery and takeaway basically rely on human transportation. The former realizes unmanned transportation, and is generally suitable for flat and clean indoor environments; the latter requires a lot of manpower and material resources, and the transportati...

AI Technical Summary

Problems solved by technology

The former realizes unmanned transportation, and is generally suitable for flat and clean indoor environments; the latter requires a lot of manpower and material resources, and the transportation cost is high

Although the former can realize unmanned transportation, it has strict requirements on the site and needs to be equipped with some auxiliary guidance equipment (magnetic strips, ribbons, reflectors, etc.), the construction period is long, and the investment cost is high.

[0004] At present, most driverless cars use laser radar as a navigation detection device, but the cost of laser radar is too high, and the extracted feature information is sparse, which is not conducive to the understanding and perception of the scene

Images