基于GPS和陀螺仪的水稻插秧机自动导航设计

基于GPS和陀螺仪的水稻插秧机自动导航设计

Computers and Electronics in Agriculture43(2004)223–234 Autonomous guidance for rice transplanting using global positioning and gyroscopes

Yoshisada Nagasaka?,Naonobu Umeda,Yutaka Kanetai,

Ken Taniwaki,Yasuhiro Sasaki

Department of Farm Mechanization and Engineering,National Agricultural Research Center,

3-1-1Kannondai,Tsukuba,Ibaraki3058666,Japan

Received in revised form20December2003;accepted30January2004

Abstract

There is an increasing trend in Japan towards consolidating,and thus enlarging,paddy?elds. Because this demands more ef?cient operations,the authors developed an automated six-row rice transplanter.It employs a real-time kinematic global positioning system(RTKGPS)for precise posi-tioning,?ber optic gyroscope(FOG)sensors to measure direction,and actuators to control steering, engine throttle,clutch,brake,etc.The RTKGPS achieves2cm precision at10Hz data output rate,and the FOG sensors are employed to maintain vehicle inclination.RTKGPS position data,in?uenced by vehicle inclination,are corrected by the FOG sensor data.FOG sensor drift is corrected by referring to the position data.To eliminate the in?uence of drift,deviation from the desired path,calculated from yaw angle and vehicle speed,is?rst compared with the deviation calculated using GPS data. Heading angle drift is then calculated.An experiment was conducted two days after?ooding the ?eld.The authors used a simple proportional steering controller.Root mean square deviation from the desired straight path after correcting for the yaw angle offset was approximately5.5cm at a speed of0.7m/s.The maximum deviation from the desired path is less than12cm which does not include the?rst2m after starting operation.Transplanting a10m×50m?eld was completed in15min.The autonomous operation was accurate enough for the rice transplanting.However,the authors could not obtain enough accuracy for spraying or mechanical weeding operations after rice transplanting because the vehicle must travel between the crop rows.For this to occur,it is necessary to improve the turn control algorithm and steering controller to obtain more precise operations.

©2004Elsevier B.V.All rights reserved.

Keywords:Automated operation;Paddy?elds;Rice transplanter;RTKGPS;FOG

?Corresponding author.Present address:Department of Agricultural and Biological Engineering,University of Illinois at Urbana-Champaign,1304West Pennsylvania Avenue,Urbana,IL61801,USA.

Tel.:+1-217-333-9420;fax:+1-217-244-0323.

E-mail addresses:nagasaka@uiuc.edu,zentei@affrc.go.jp(Y.Nagasaka).

0168-1699/$–see front matter©2004Elsevier B.V.All rights reserved.

doi:10.1016/pag.2004.01.005

基于GPS和陀螺仪的水稻插秧机自动导航设计

224Y.Nagasaka et al./Computers and Electronics in Agriculture43(2004)223–234

1.Introduction

With farm size increasing and the number of farmers decreasing,more ef?cient agricul-tural practices are needed.However,in Japan,?elds are not as yet suf?ciently consolidated. They are frequently dispersed and it is dif?cult to use large machines in areas that would be unable to bear their weight.This ef?ciency problem would be solved,however,if one operator were able to control several lightweight machines simultaneously.The goal of our research is to develop an autonomous operating system for paddy?elds that would permit one operator to control multiple machines.The objective of this study is to develop an autonomous navigation system for the rice transplanter.

Several pieces of research on autonomous agricultural machinery have been published (Keicher and Seufert,2000;Reid et al.,2000).Recently,an autonomous tractor using an optical surveying device and a terrestrial magnetism sensor for ploughing was developed (Yukumoto and Matsuo,1995;Masuo et al.,2002).They obtained good results but found it dif?cult to control multiple machines because each vehicle required its own optical sur-veying device.Some researchers have developed machine vision-based vehicle guidance system(Pilarski et al.,1999;Han et al.,2002).Although it was effective for vehicle control along detected crop rows,this type of guidance system is not useful for rice transplanting because transplanting operations takes place in?ooded?elds and the re?ection of sunlight on water interferes with image processing.Some researchers used differential GPS(DGPS) (Ramalingam et al.,2000;Cho and Lee,2000).Kalman?ltering of DGPS can effectively correct DGPS position error(Han et al.,2002),but it does not provide enough precision for the operation in paddy?elds.Researchers at Stanford University(O’Connor et al.,1996; Bell,2000)used carrier-phased DGPS and obtained good results.Noguchi et al.(2002a,b) developed a robot tractor and Nagasaka et al.(1997,2000)have developed an automated rice transplanter.Both of them employed real-time kinematic global positioning system (RTKGPS)and?ber optic gyroscope(FOG)sensors that established precise operations.In this paper,the results of automated rice transplanting operation experiments are reported. The authors?rst describe the sensors,actuators and controllers used for the automated rice transplanter and how data processed.Then The authors describe the vehicle control methods and how the desired operating path is chosen.The experimental results are discussed,and the?nal section gives conclusions.

2.Materials and methods

2.1.Vehicle,sensors,actuators,and controllers

The authors modi?ed a commercial six-row rice transplanter(PH-6,Iseki Co.,Ehime, Japan).Fig.1shows the modi?ed rice transplanter controlled by a computer with a486-compatible66MHz central processing unit.Fig.2is a schematic for an automated rice transplanting system.An RTKGPS(MS750,Trimble Navigation Ltd.,Sunnyvale,CA)with 2cm precision at10Hz data output was used to locate the position of the rice transplanter. For communication between the RTKGPS reference station and the rover station,5-mW output wireless radio modems(YRM211,Vertex Standard CO.,Ltd,Tokyo,Japan)were

基于GPS和陀螺仪的水稻插秧机自动导航设计

Y.Nagasaka et al./Computers and Electronics in Agriculture43(2004)223–234225

Fig.1.Automated rice transplanter.

used with baud rates set at9600bps.A FOG sensor(JG-35FD,Japan Aviation Electronics Ind.Ltd.,Tokyo,Japan)was used to measure the yaw angle,and an inclination-measuring apparatus(JCS-7401,Japan Aviation Electronics Ind.Ltd.)comprised of three FOG sensors and three accelerometers was used to measure the roll and pitch angles.Table1shows the FOG sensor speci?cations and the inclination-measuring apparatus.

The GPS position data,yaw,roll and pitch angle data were transferred to a computer through an RS232C interface.The sampling rate was10Hz for the GPS,20Hz for the yaw angle,and25Hz for the roll and pitch angles.A one-chip-microcomputer generated10Hz output through the RS232C interface and triggered the sampling for the GPS,the FOG and

Fig.2.Schematic of an automated rice transplanting system.

基于GPS和陀螺仪的水稻插秧机自动导航设计

226Y.Nagasaka et al./Computers and Electronics in Agriculture43(2004)223–234

Table1

FOG sensor and inclination-measuring apparatus speci?cations

JCS-7401JG35-FD Resolution(?)0.20.01 Measurement range(?)±45±180 Accuracy(?)0.2–

Drift(?/h)–20 Linierity(%)10.1 the inclination-measuring apparatus data.The main computer corrected the GPS position data that was in?uenced by the vehicle inclination,and calculated the control parameters, then sent them to an industrial programmable logic controller(PLC)(KZ-350,Keyence CO.,Osaka,Japan)through the serial port every100ms.The PLC receives control parame-ters from the main computer through the RS232C interface link unit to control the actuators. Fig.3shows the actuators that control the levers,pedals,etc.All actuators are connected to relays,allowing direction of operation and speed to be controlled by pulse width from the PLC.The steering angle is sensed by means of an absolute rotary encoder and is controlled by a dc motor.Proximity sensors detect the clutch and brake positions and electrical linear cylinders control the clutch and brake pedals.The positions of the transplanting instrument

Fig.3.Sensors and actuators.

基于GPS和陀螺仪的水稻插秧机自动导航设计

Y.Nagasaka et al./Computers and Electronics in Agriculture43(2004)223–234227

Fig.4.Correction of inclination.

control levers,engine throttle and the hydro static transmission(HST)lever were measured using absolute rotary encoders and operated by electrical linear cylinders.The PLC control loop was2ms.

2.2.Correction of inclination

When a rice transplanter travels in a paddy?eld,the unevenness of the ground causes roll and pitch angles of about3?.In this system,the GPS antenna is attached to the rice transplanter.The GPS data indicate the position of the top of the GPS antenna.Assuming the antenna height to be2m,the horizontal distance between the top of the antenna and the bot-tom is10.5cm when the roll angle is3?.To determine the exact position of the ground surface point under the GPS antenna,the in?uence due to roll and pitch angles must be corrected. Fig.4shows the correction method due to inclination.Assuming the presence of space coordinates xyz on the rice transplanter and plane coordinates XYZ on which the rice trans-planter travels,the straight desired path was derived on XY.In Fig.4,P is the GPS antenna-tip coordinate,θis the roll angle,φis the pitch angle,the yaw angle isψand the GPS antenna height is h.The coordinates of point P after being corrected is given by P0.P0is expressed by the following equation,in which R(?φ),R(?θ),and R(?ψ)are the rotation around the X,Y,and Z axes.

P0=R(?φ)R(?θ)R(?ψ)P(1) 2.3.Correction of initial yaw angle offset and FOG drift

It is dif?cult to set the vehicle direction parallel to the traveling direction at the starting point of operation.The FOG sensor cannot sense the azimuth,and thus,suffers drift.To sense the initial yaw angle and to correct for FOG drift,the deviation of lateral direction calculated using the yaw angle and vehicle speed was compared with the deviation calcu-lated using GPS data(Fig.5).The offset and drift angle were then calculated and the yaw angle corrected.

For the short time interval, t,the deviation of lateral direction d can be calculated using Eq.(2).In this equation,v is vehicle speed and the yaw angleψis small.

d=v sin(ψ) t～=vψ t(2)

基于GPS和陀螺仪的水稻插秧机自动导航设计

228Y.Nagasaka et al./Computers and Electronics in Agriculture 43(2004)

223–234

Fig.5.Correction of initial yaw angle offset and drift.

The deviation of lateral direction d is obtained by integrating both sides.d =

vψd t (3)The deviation of lateral direction for a de?nite period of time measured by the RTKGPS is d GPS .For the discrete data value,the deviation of lateral direction d Gyro estimated by the vehicle speed and the yaw angle is expressed as Eq.(4).In this equation,the yaw angle at a point of time i is ψ(i ),the vehicle speed,measured by GPS,is v (i ),n is the time after starting the calculation and t s is the sampling interval.

d Gyro =n/t s

i =0v(i)ψ(i)(4)

When the distance is l n ,the offset of yaw angle ψoffset is calculated as Eq.(5).In this equation,the authors assume that d GPS and d Gyro are suf?ciently smaller than l n .ψoffset =arctan d Gyro l n ?arctan d GPS l n ～=d Gyro ?d GPS l n

(5)d Gyro is calculated at a point 35–40m after starting operation.The condition of the paddy ground is ?at in this interval,and thus,stabilizes the motion of the vehicle.If the interval is longer,the calculated results are in?uenced by the vehicle sideslip and the error is inte-grated.So,in this study,l n is 5m and ψoffset is calculated at a point 35–40m after starting operation.The initial offset and the FOG sensor drift were corrected for every operation path before starting.

2.4.Path planning

Before starting operation,the computer must create a desired path,along which the rice transplanter travels.In this study,the paddy ?eld was assumed to be rectangular.The four corners:A,B,C,and D,in the ?eld were measured beforehand.The sowing width of a six-row rice transplanter is 1.8m.Since a paddy ?eld usually has only one entrance,the desired path depends on the width of the ?eld.

When the quotient (the ?eld width divided by the sowing width of 1.8m)is an even number,the automated rice transplanter starts operating near the entrance.The desired path is set up so that the distance from side AB to the ?rst path and that from side CD to the last path are equal.When the quotient is an odd number,it starts operating at the opposite

基于GPS和陀螺仪的水稻插秧机自动导航设计

Y.Nagasaka et al./Computers and Electronics in Agriculture 43(2004)223–234

229(a)(b)

C

C

Fig.6.Planed desired path (the quotient is the ?eld width divided by the operation width).(a)The quotient is even;(b)the quotient is odd.

side of the entrance.The desired path is set up so that the distance from side CD to the ?rst path and that from side AB to the last path are equal.While the vehicle is moving to the operation starting point,the moving path is set at a distance of 30cm from the last desired operating path.Fig.6shows the desired path planned before operation.

2.5.Straight drive

In paddy ?elds,it is thought that the paddy ground is more ?imsy than upland ?elds.There are several papers about steering control in upland ?elds.O’Connor et al.(1996)used a steering controller based on a set of linear motion equations.Cho and Lee (2000)used a fuzzy controller for an autonomous operation for an orchard speed sprayer.Kise et al.(2002)developed an optimal steering controller and they obtained good curved-path guidance results.Inoue et al.(1997)developed an adaptive steering controller correcting steering system delay.Most of them analyzed vehicle dynamics and made a dynamic model,then decided the control parameters.

However little research has been done regarding vehicle dynamics (Weise et al.,2000)in paddy ?elds.The authors could not make a dynamic model at this time due to the soil conditions which were not constant,making it is dif?cult to ?x the parameters.So in this study,the authors used a simple proportional controller for steering control before evaluating effectiveness of the advanced controllers in such loose-ground like paddy ?elds.

The rice transplanter must be driven along the desired path.The steering is set to follow as closely as possible to the desired path.The steering angle δaim is given as Eq.(6),where the deviation from the target line is d and the yaw angle is ψ.The value of K p1,K p2is determined according to the vehicle speed.

δaim =K p1d +K p2ψ(6)In this study,the steering control parameters were obtained experimentally.K p1is 0.636and K p2is ?0.604at a speed of 0.7m/s.The steering speed is proportional to the difference between the target steering angle and actual angle,and is controlled by the pulse width mod-ulation (PWM).The GPS data quality indicator was monitored while the rice transplanter traveled automatically.If the radio link between the GPS base station and GPS rover station was lost for any reason,the clutch is released and the operation is interrupted.The engine

基于GPS和陀螺仪的水稻插秧机自动导航设计

230Y.Nagasaka et al./Computers and Electronics in Agriculture43(2004)223–234

Fig.7.Control way of turning.

throttle,the HST lever,and the attachment control lever positions were controlled according to the vehicle location.The engine throttle position is set fully open during operation and is set at the middle position in other situations.The HST lever position is set so that the vehicle travels at a speed of0.7m/s.

2.6.Turn control

At the headland,the rice transplanter moves forward and backward during a turn so as to minimize the headland space.The width of the headland is3.5m.Fig.7shows how turning is effected at the headland.When the rice transplanter reaches the edge of the?eld,it?rst stops transplanting and lifts the transplanting attachment,then the HST lever is set to the backup position and the vehicle moves backward in a straight line for a distance of40cm.After mov-ing40cm,the HST lever is set to the advance position at0.3m/s.While the rice transplanter is turning,if the yaw angle is less than160?,only the yaw angle is obtained.Under these con-ditions,the steering angle is kept at40?and the brake is applied on one side.If the yaw angle exceeds160?,the rice transplanter is moved as close as possible to the next desired path.The transplanter stops4.7m from the?eld edge.Then it moves backward.If the rice transplanter does not get suf?ciently close to the next desired path at this point,the steering is manipulated so as to get as close as possible to the desired path until it reaches2m from the?eld edge.

3.Results and discussion

The experiment was conducted two days after?ooding the?eld.The four corners of the paddy?eld were measured before operation.The size of this paddy?eld was50m×10m. The quotient obtained by dividing the?eld width by the sowing width was an odd number, so in this?eld,the automated rice transplanter moved to the opposite side of the entrance before starting operation.

Fig.8shows the path of the automated rice transplanter.In Fig.8,the in?uence of the vehicle inclination was corrected.Table2shows the summary of line-following control at a speed of0.7m/s.

The rice transplanter went forwards and backwards in the?eld three times.Path1has an RMS position error that is more than0.2m.It is thought that the?rst yaw angle offset was large.The lateral position error in Table2(a)is larger than in(b)because in the?rst

基于GPS和陀螺仪的水稻插秧机自动导航设计

Y.Nagasaka et al./Computers and Electronics in Agriculture 43(2004)223–234231

Fig.8.The path of the automated rice transplanting operation.

Table 2

Summary of line-following control

Path 1

Path 2Path 3Path 4Path 5(a)Includes entire operating path Path width (m)

0.3480.1530.3020.1940.195Lateral position from the desired path,mean (m)?0.2420.018?0.0150.031?0.098Standard deviation (m)0.0590.0400.0840.0450.029Lateral position RMS 0.2420.0410.0700.0420.098Heading error,mean (rad)

0.0220

?0.04330.0131?0.04770.0290(b)Not including the ?rst 2m after starting operation Path width (m)0.1660.1490.2200.1560.136Lateral position from the desired path,mean(m)?0.2550.0220.0210.025?0.095Standard deviation (m)0.0370.0380.0530.0410.026Lateral position RMS 0.2550.0410.0490.0360.095Heading error,mean (rad)0.0251

?0.0332

0.0044

?0.0421

0.0185

2m after starting operation the transplanter was not close enough to each desired operating path.Fig.9shows the path of the rice transplanter traveling on concrete.In Figs.8and 9,Although the turning radius of the vehicle at the headland was around 2m and it was easy to get close to the new desired path after turning,the direction of the vehicle before starting operation was not close enough to the desired direction.In Fig.8,when the vehicle went backward,it traveled along the turning path.This result was not obtained on concrete.It was thought that the vehicle traveled along its wheel track and did not get close enough to the desired path at the beginning of the operation.One way in which we solved this problem was to move the position of the limit switch and increase the force on the brake pedals,giving a smaller turning circle and making it easy to point the vehicle in the desired direction.Another way is to have the vehicle enter every other or every second operation

Distance(m)

50

D i s t a n c e (m )

Fig.9.The path on concrete.

基于GPS和陀螺仪的水稻插秧机自动导航设计

232Y.Nagasaka et al./Computers and Electronics in Agriculture 43(2004)223–234

D e v i a t i o n (m )

Y a w a n g l e (r a d i a n )

S t e e r i n g a n g l e 5

45

0.10

Travel distance(m)

(r a d i a n )

Fig.10.The steering angle,the yaw angle and deviation in operation.

path.In this case,the turning circle is large,so it is possible to turn without braking.It is necessary to improve and to better evaluate the turn control algorithms.Doing this might make it possible to reduce the backup distance and to save turning time.

Fig.10shows the steering angle,the yaw angle and the deviation for the third path of the automated rice transplanting operation.Fig.11shows those parameters for traveling on concrete.In Fig.10,the steering angle ranges from ?0.08to 0.04.The yaw angle ranges from ?0.07to 0.The deviation ranges from ?0.07to 0.08m.In Fig.11,the steering angle,the yaw angle and the deviation are almost constant.In Fig.10,the steering angle is 0between 25and 45m from the starting point,even though the yaw angle is shifted by ?0.04and the deviation is shifted by 0.04m.This is due to insuf?cient accuracy even though the FOG was set using the calculated result by Eq.(5).In paddy ?elds,because of ?imsy ground and variable terrain conditions,sideslip sometimes occurs during the calculation of the yaw angle offset.In this study,the steering control parameters were obtained experimentally.The authors thought that it would be dif?cult to estimate the location and direction of the vehicle by relying only from calculation of vehicle dynamics,since soil conditions are not constant and

5

45

D e v i a t i o n (m )

Y a w a n g l e (r a d i a n )

S t e e r i n g a n g l e Travel distance(m)

(r a d i a n )

0.10

-0.10.10

-0.1

0.10-0.1

Fig.11.The steering angle,the yaw angle and deviation on concrete ground.

基于GPS和陀螺仪的水稻插秧机自动导航设计

Y.Nagasaka et al./Computers and Electronics in Agriculture43(2004)223–234233 paddy?elds are innately slippery.However,as can be seen in Fig.10,there is a correlation between steering angle and yaw angle,which leads to deviation from the desired path.GPS data have a latency of about70ms and the automated system also has a latency that is a function of mechanical delay,calculation time and data transfer time.It should,therefore, be possible to improve the precision of the operation if the automated rice transplanter is controlled based on vehicle dynamics on soft ground.

In this experiment,conventional mat-type rice seedlings were used,with two persons supplying seedlings to the rice transplanter at the edge of the?eld.At the headland,it took about60s to turn and realign to the next desired path.Planting the experimental plot took 15min.

4.Conclusion

An automated rice transplanter was developed and autonomous guidance operation was conducted.In this study,the position data that was in?uenced by the inclination of the vehicle was corrected by FOG sensors data.FOG drift was corrected by comparing the deviation of lateral direction calculated using the yaw angle and vehicle speed with the deviation calculated using GPS data.The maximum deviation from the desired straight path is12cm and deviation is5.5cm at0.7m/s operating speed after correcting for the yaw angle offset.This is only enough accuracy for rice transplanting operations,and it is not accurate enough for spraying and mechanical weeding operations after rice transplanting. Since it is necessary to drive an autonomous vehicle between the crop row to make spraying and mechanical weeding operations,the maximum deviation must be reduced less than5cm. In this study,a simple proportional controller was used for steering control.However,for more precise operations,it is necessary to analyze the vehicle dynamics in the paddy?elds and to evaluate and improve the steering controller.Since the manual-operating speed of a rice transplanter is usually around1m/s,it is also necessary to evaluate the control algorithm at speeds of more than1m/s.

At the headland,there was more than a10cm deviation from the desired path at the beginning of the operation.As the vehicle traveled along the wheel track when moving backward,the turning circle diameter must have been the same as the distance to the following operating path.It is necessary to evaluate the turn control algorithms.

Rice transplanting operation requires someone to supply the seedlings.It was necessary to supply seedlings every200m in this experiment.This poses a problem for making ef?cient operations and for establishing fully autonomous rice transplanting operations in larger paddy?elds.This,however,is solved when long mat type hydroponic seedlings (Tasaka,1998)are used.This would require a seedling supply every30ares for a six-row rice transplanter so more ef?cient operations could be made.

Acknowledgements

This research project was supported by the Ministry of Agriculture,Forestry and Fisheries of Japan.

基于GPS和陀螺仪的水稻插秧机自动导航设计

234Y.Nagasaka et al./Computers and Electronics in Agriculture43(2004)223–234 References

Bell,T.,2000.Automatic tractor guidance using carrier-phase differential puters and Electronics in Agriculture25(1/2),53–66.

Cho,S.I.,Lee,J.H.,2000.Autonomous speed sprayer using differential global positioning system,genetic algorithm and fuzzy control.Journal of Agriculture Engineering and Research76,111–119.

Han,S.,Dickson,M.A.,Ni,B.,Reid,J.F.,Zhang,Q.,2002.A robust procedure to obtain a guidance directrix for vision based vehicle guidance system.in:Proceedings of the Automation Technology for Off-road Equipment, ASAE,Chicago,IL,pp.317–326.

Han,S.,Zhang,Q.,Noh,H.,2002.Kalman?ltering of DGPS position for a parallel tracking application.

Transaction of ASAE45(3),553–559.

Inoue,K.,Otuka,K.,Sugimoto,M.,Murakami,N.,1997.Estimation of place of tractor and adaptive control method of autonomous tractor using INS and GPS,Preprints of the International Workshop on Robotics and Automated Machinery for Bio-Productions,EurAgEng,Gandia,pp.27–736.

Keicher,R.,Seufert,H.,2000.Automatic guidance for agricultural vehicles in puters and Electronics in Agriculture25(1/2),169–194.

Kise,M.,Noguchi,N.,Ishii,K.,Terao,H.,2002.The development of the autonomous tractor with steering controller applied by optimal control.in:Proceedings of the Automation Technology for Off-road Equipment, ASAE,Chicago,IL,pp.367–373.

Masuo,Y.,Yamamoto,S.,Yukumoto,O.,2002.Development of tilling robot and operation software.in: Proceedings of the Automation Technology for Off-road Equipment,ASAE,Chicago,IL,pp.184–189. Nagasaka,Y.,Otani,R.,Shigeta,K.,Taniwaki,K.,2000.A Study About an Automated Rice Transplanter with GPS and FOG.ASAE Paper No.00–1045.ASAE,Milwaukee,WI.

Nagasaka,Y.,Taniwaki,K.,Otani,R.,Shigeta,K.,1997.Automated operation in paddy?elds with a?ber optic gyro sensor and GPS.Preprints of the International workshop on Robotics and Automated Machinery for Bio-Productions,EurAgEng,Gandia,pp.21–26.

Noguchi,N.,Kise,M.,Ishii,K.,Terao,H.,2002a.Field automation using robot tractor.in:Proceedings of the Automation Technology for Off-road Equipment,ASAE,Chicago,IL,pp.239–245.

Noguchi,N.,Will,J.,Ishii,K.,Reid,J.F.,2002b.Development of master–slave system-obstacle avoidance algorithm.in:Proceedings of the Automation Technology for Off-road Equipment,ASAE,Chicago,IL, pp.432–441.

O’Connor,M.,Bell,T.,Elkaim,G.,Parkinson,B.,1996.Automatic Steering of Farm Vehicles Using GPS.Paper presented at3rd International Conference on Precision Agriculture,ASA,CSSA,and SSSA,Madison,WI. Pilarski,T.,Happold,M.,Pangels,H.,Ollis,M.,Fitzpatrick,K.,Stentz,A.,1999.The DEMETER system for autonomous harvesting.in:Proceedings of the8th International Topical Meeting on Robotics and Remote Systems,ANS,Pittsburgh,PA.

Ramalingam,N.,Stombaugh,T.S.,Mirgeaux J.,2000.DGPS-Based Automatic Vehicle Guidance.ASAE Paper No.001068,ASAE,Milwaukee,WI.

Reid,J.F.,Zhang,Q.,Noguchi,N.,Dickson,M.,2000.Agricultural automatic guidance research in North America.

Computers and Electronics in Agriculture25(12),155–167.

Tasaka,K.,1998.Outline of raising and transplanting technology for long mat type hydroponic rice seedlings.

Ag.Eng.Paper98-A-061.EurAgEng,Oslo.

Weise,G.,Nagasaka,Y.,Taniwaki,K.,2000.An investigation of the turning behaviour of an autonomous rice transplanter.Journal of Agriculture Engineering and Research77(2),233–237.

Yukumoto,O.,Matsuo,Y.,1995.Research on autonomous land vehicle for agriculture.in:Proceedings of the International Symposium on Automation and Robotics in Bioproduction and Processing,vol.1,JSAM,Kobe, pp.41–48.

本文来源：https://www.bwwdw.com/article/wvaj.html