| Harvesting transportation is one of the most common working cycles during vegetable producing inside greenhouse,which calls for high workload and frequent operation.Properly utilizing automated handling tools will help to improve operational efficiency,reduce working burden and promote the development of modern agriculture.The currently used railway cart in some way helps to solve this problem.However,in China,most of the greenhouse is sunlight greenhouse,which is not suitable for the railway vehicles.So in this paper,considering the transportation problem inside greenhouse and the current greenhouse structure character in China,a kind of following system designing method for autonomous transportation was presented.This system was designed to realize automatic following after a workingman inside the greenhouse,which would help to deal with the transportation during work.An intelligent control system for the automatic following system was designed,and experiments were carried out to confirm the reliability of the automatic following system.The main research contents and conclusions are as follows:(1)A human body checking system was designed based on Kinect sensor.As the vehicle was mainly used inside greenhouse after the workingman,a navigation and positioning system was designed with the Microsoft sensor,which could perceive distance.A lot of algorithms such as weighted median filter,image segmentation and edge extraction were introduced to achieve body skeleton from depth.picture,thus obtaining the target coordinate.The relevant experiment showed that the observed data had significant correlation with the true data,indicating reasonable model for later designing.(2)The automatic navigation system was established.According to the working requirements and environment in greenhouse,mechanical structure was designed and manufactured.Kinect was set at the very height;vehicle structure parameters,load performance and steering performance were also analyzed.To make it more comfortable for workers,human factors engineering was considered to perfect the system.(3)An intelligent control system for the automatic following system was designed.Pure pursuit method was introduced to complete the tracking while fuzzy control was used to determine the look-ahead distance.As the trajectory inside greenhouse was complex and unknown,adaptive function was applied to optimize the path firstly.At the depth direction,if the distance between the vehicle and the workingman was beyond the threshold for a very short time,the system didn’t start,but it would be regarded as starting signal when the time was not very short.Similarly,at the lateral direction,if the trajectory changed greatly in very short time,it didn’t start,vice versa.(4)The straight line tracking experimental results indicated that the largest depth deviation was 8.5cm.The designed controlling system and the original system both had a good performance when the road was smooth;but compared with the original controlling system when the road changed unexpectedly,the designed system was able to turn freely and precisely while avoiding abrupt turning.The working function had nothing to do with the light condition,which could meet the demands of harvesting and goods transportation in greenhouse in different seasons.(5)To further determine the system performance,test under different load,different road conditions and different guiding speed were carried out.The results were as follows:a.The system showed good performance within limited load,the largest following time shorter than 10S,with deviation smaller than 10cm.b.Compared with smooth road,following time under rough road and road with branches and leaves was significantly long,with little difference on depth deviation.c.The following signal lost when the accelerated speed of the workingman reached.As speed of vehicles inside greenhouse was usually low,accelerated speed had little influence on the system. |
PDF Full Text Request