GB2202650A - Automatically piloted vehicle - Google Patents

Description

LOCOMOTION-COMMAND METHOD FOR MOBILE ROBOTS BACKGROUND OF THE INVENTION This invention relates to a method for giving a locomction-command system for obile robots such as unmanned travelling vehicles.

There have been proposed various methods for raking unmanned mobile robots travel; one nethod comprises the steps of emitting inductive radio waves of a predeternined frequency fro an induction wire and making a mobile robot receive the radio waves to travel along the route formed by the induction wire. Another method uses an optical reflective tape and a photoelectric detector instead of the induction wire. These methods require guiding means, as the mobile robot of this type are controlled and guided by detecting and coipensating deviation frow the route of the induction wire or the photoelectric reflective tape which is laid on a floor.When the layout of tools and machineries and/or facilities in a plant or a warehouse are revised, the conveyor route formed arans such facilities should be changed accordingly, which inconveniently needs replacing the guiding isans again.

Further in the case of the method using an induction wire, it takes much tile and trouble in laying and burying the wire. The method using a photoelectric reflective tape has another disadvantage that it is prone to dust and easily stained or daaged.

Japanese patent laid-open No. 62424/1982 discloses a travel-command method which obviates aforesentioned detriments In the method. the route along which a mobile object is to travel is erpressed by a sequence of points with X-Y coordinates, which are generated in a for, of digital data, with

@@ a a region is determined. The region and an assured position ahead in the advancing direction deteriine a region to which the object is to advance. A steering signal is generated based upon a point on a line connecting at least two points included within the specified region and the assured position of the moving object ahead thereof.However, as the advancing route is expressed with a sequence of points with coordinates in this method, it is not possible to smoothly specify or to control the route when the travel-lane should change fros a line 1 to another parallel line 2 as shown in FIG. 1A, or when the object should sake a U-turn and return on a sare route 3 as shown in FIG.1B.

SUMMARY OF THE INVENTION An object of this invention is to provide a method for giving commands to mobile robots which enable the robot to freely travel without laying inductive wires or photoelectric reflective tapes on a floor.

Another object of this invention is to provide a method for giving locomotion-commands which enable an object to freely travel along an arbitrary route, which can easily change travelling route, and which can easily @ake the object mke a U-turn.

Still another object of this invention is to provide a method for giving locomotion-commands which enable a mobile robot to travel freely with a simple leans and which is capable of controlling communication between a master and the robot effectively According to this invention in one aspect thereof, for achieving objects described above, there is provided a locomotion-command nethod for a mobile robot of the type having a aster and a locomotion nodule where a travelling route is specified by a cos and sent fron the aster to the @@@omotion nodule, which is characterized in that travelling on a given directed line is to be a basic notion, and a travelling route can be arbitrarily specified by sending a command which changes on the position and direction of said directed line.

According to this invention in another aspect thereof, there is provided a locomotion-command method for a mobile robot of the type roving while carrying out co ication between a aster and a locomotion nodule which is characterized in that the feedback control tine during which the robot per se executes travelling control and the con and execution control tine during which commands fro said aster are executed and controlled are alternately assigned for a predetermined tine.

Further, according to this invention in still another aspect thereof, there is provided a loconotion-conind method for a mobile robot of the type which moves the robot while communicating between a aster and a locomotion nodule, which is characterized in that said locomotion nodule has a feedback control node by which the loconotion nodule control its travel by itself and a command execution control node by which coneands froi said aster are executed and controlled, the position and direction on the travelling line of said locomotion nodule are specified by using coordinate transformation as said command, the given command and said comand are analysed in the execution control node, and the travel of the robots controlled according to the result of said analysis in said feedback control rode.

The nature, principle and utility of the invention will become nore apparent froi the following detailed description when read in conjunction with the accompanying drawigngs.

BRIEF DESCRIPTION OF THE DRAWINGS In the attached drawings: FIGS. 1A and 1B are the views to show examples of the movement of a robot; FIGS.2 and 3 are the views to explain the principle of this invention, respectively; FIG.4 is a block diagra, to show the relation between a aster and a locomotion module of a robot according to this invention; FIG.5 is a diagram to show the communication between the aster and the locomotion nodule; FIG.6 shows the apperarance of an embodiment of the mobile robot according to this invention; FIG.7 is a block diagras to show an embodiment of the controlling sister thereof; FIG.8 is a graph to show the timing of movement thereof; FIGS 9 through 11 are flow charts to show respective movements of the mobile robots; ; FIGS. 12A through 12H are the views to show the movement of a mobile robot whith the coneand of "CO, respectively; and FIG. 13 is a view to show the state of travel of a mobile robot.

DESCRIPTION OF THE PREFERRED EMBODIMENT The principle of this invention will now be described referring to FIG. 2.

In FIG.2, a robot R (x,y) travels on a directed travel line X in the direction marked with an arrow. The mobile robot R is controlled positionwise servo nechanisn so that it would not deviate from the line X by a large margin although it may go slightly toward the direction of Y-uis. In other words, the robot R is constantly servo-controlled to neke the relation y=O.

When one wishes to have the robot neke a U-turn or change direction, the commanding section gives a new directed line X' and an exit point E (e,o) which starts the transient state fron the line X to the line X' is specified as a transition point. With this command, the robot R is controlled to travel at a given speed until it reaches the transition point E and then to transfer to a new route X'. The travel line X is an X-axis of the XY-coordinates with a fixed original point 0 (0,0). When x # e, the robot R is controlled to enter the transient state and is transferred to a new X'Y'-coordinate baring an origin 0' (0,0) from a point on the X'-axis. The point is herein referred to as F.The robot R has a position (x,) and a direction 8 in itself until the time it reaches the transient point E and constantly renews the current values with an encoder. The direction 8 is an angle formed between the axis X and the mobile robot R.The direction is controlled so that a target value is set at y=0 under stationary state wherein the direction CCW is positive and posture of the robot R is set at (x,y,#). After the transient point E, the posture of the robot R is controlled cm a new coordinate X'Y' as (x',y',#',) The coordinate transfornation fron the Xf-coordinate to X'Y'coordinate is described below referring to FIG. 3. It is assured when a command "GO" (which will be explained hereinafter) specifies a new coordinate X'Y' to the old coordinate XY, the posture of the robot R which was (x,y, #) is (x', y', #') in a new coordinate.The relation holds between the above two as below;

Froi the above formula (1), the following relations are induced;

When the command "GO" and so on are executed, the coordinate transformation of the formula (2) or the transformation from (x,y,#) to (x',y',#') is carried out.

FIG.4 shows the relation between the aster 10 and the others including the locomotion nodule 20 of a mobile robot. A vision nodule 11 provides a sensor which identifies surroundings and a hand-and-ar. nodule 12 provides a anipulator. FIG.5 shows that communication data comprising commands C and replies RP is exchanged between the locomotion nodule 20 and the aster 10 at a tinting described hereinafter. Commands C are transmitted froi the aster 10 to the locomotion nodule 20 in the unit of one byte. A separator such as a code or a symbol ";" is inserted between coinds.

Conands are classified into a slow command "SC" which is temporarily stored in a buffer nenory BM and then executed sequentially in the order of arrival and a fast cos and "FS" which is immediately executed. Slow commands "SC" are temporarily stored in a buffer memory BM and then executed in a first-in first -out basis. Fast commands "FC" are executed innediately without being stored in a buffer memory Bn. The locomotion nodule 20 has two states; a wait state and an active state. In the wait state, a commend stored in the buffer memory BM is not executed.The state is actuated to an active state when a fast command or a "START" cos and t which will be described hereinafter ) is transtitted fron the aster 10. The locomotion nodule 20 is in the waiting state when initialized. In the active state, the locomotion module 20 executes commands stored in a buffer memory BM in the order of arrival. Execut ion of each coy and C is coipleted by the tile the step reaches a transient point of a subsequent command for instance C1 ) where the execution of the command C1 starts.If there is no command stored in a buffer memory BM during the tine a co CO is being executed, or if the buffer memory BM is empty, the execution of the cos and CO continued indefinitely. However, it is always

p#ble to end the execution of the command CO by transmitting such commands as "Go", "STOP 0" or "ADJUS. An active state turns into a waiting state when a con and "WAIT" r ( descirbed hereinafter ) is executed. This con and "WAIT" is a slow command which is accepted only during the active state.

Replies RP are classified into "POSTURE" indicating data for the posture of the robot, "TRANSIENT" indicating that the robot has passed the transient point E but not reached the point F of the next coordinate, "STATIONARY" indicating the state is stationary, and "ERROR" indicating that an exception has taken place, that is, the state will not return to the stationary state for a long tile.

An example of control algorithm is shown below for executing commands such as "GO", "SPIN", "STOP", "ADJUST", or "STOP 0". The velocity V and the angular velocity # are servo-controlled with a program and data stored in ROM ( Read Only We orgy ) or RAN ( Random Access Memory ) and by the CF ( Central Processing Unit ) to follow up the target values.

When a command "GO" is given to a robot specifying the target position xd as the point to stop, the velocity is controlled at a fixed velocity node until it becomes a predeternined state, and then at a reduced velocity node until the X becomes a predeternined value. Then, the node is switched to a position controlling node, and finally to a suspension node. In order to execute the cox and, a target velocity Vr(t) is given to correspond to the current position x(t) for speed reduction, and when the actual speed and the position satisfy V(t)# 0, and t(t) xd, the servo-loop is switched fron a velocity servo-cantrol node to a position servo-control node.For advancing a robot at a given speed until the target position xd and stopping

#same tbereon, the velocity V(x) at the point x should be;

Viz)= kl signtxd - s) lsd - - I " (3) If it is assured that the reduction in velocity is Av, the coefficient kl is expressed as

When the relation erpressed by the following formula (5) is obtained during the executin of a conid, the node goes into a velocity reduction node, wherein Z is an extrenely snall value.

Vr(t) # V( x(t) ) - # ............ (5) In the velocity reduction node, for commend controlling, the angular velocity # is controlled by changing f(r, #, #) as follows; fo(r, #, #) = k2 . # - k3 . # ............ (6) This control is continued in the position control node.The control of the velocity V is expressed as follows: Vr ( t + #t ) = - Vmax ............ (7) if Vr - At .g < -Vmax, Vr ( t + #t ) = Vmax ............ (8) if Vmax < Vr+ #t.g, Vr ( t + #t ) = Vr (t) + #t # GO x ( V(t), x(t) ) ............ (9) if -Vmax# Vr + #t.g # Vmax, wherein Go is either -Av or +Av.

After controlling in the velocity reduction node, if

are satisfied, the state goes into a position control node. It is controlled under the position control node so that Vr(t) = k4#( xd - x(t) ) ............ (11) When the relation below is satisfied after the positon control, it is judged that the robot has reached and stops at the target position xd.

The node is called a stop lode.

FIG.8 sbows the configuration of a mobile robot of powered steering system of which this invention is applied. The botton of the robot is provided with a pair of travelling wheels 101 and 102 having a width # there- between while the top thereof is provided with a ultrasonic range finder 103.

A locomotion nodule is mounted on a printed board and so on within a rack 104.

The locomotion nodule functions to analyse commands C sent from the waster 10 or drive the wheels 101 and 102 and comprises a part of the locomotion nodule 20. FIG.7 shows the structure thereof in lore detail. The loconotion nodule includes a icro-coiputer, CPU 110. The CPU 110 is connected a comminication controller 111 to coninicate with the commanding section, a ROM 112 which stores prograns and parauters, and a RAN 113 which temporarily stores the data required for controlling. The CPU 110 is connected to a controller 114L which controls the driving of the left wheel 101 and a controller 114R which controls the driving of the right wheel 102.The controllers 114L and control the driving of the wheels 101 and 102 via rotors 115L and 115R, respectively The wheels 101 and 102 are connected to shaft encoders 117L and 117R, respectively. The pulse outputs PRL and PRR fron the shaft encoders 117L and 117R are given to the controllers 1141 and 114R via the CPU 110, respectively. Brakes 116L and 116R are controlled by motors 118L and 118R are driven with brake controlling singals B from tbe CPU 110.

The CPU 110 gives the target velocities VrL and VrR to the controllers 114L and 114R. Suppose the velocity of the robot 100 is V and the angular velocity is co, then

Therefore, the target velocities VL and YR are determined bJ the following relation my be induced fron the formula (13).

If the velocity V and the angular velocity co of the robot 100 are given, the velocities VL and VR of the left and right wheels 101 and 102 are calculated according to the above formula (14) and are given to the controllers 114L and 114R as the target velocities VrL and VrR, respectively. The velocities VL and VR of the wheels 101 and 102 are measured by the shaft encoders 117L and 117R, respectively. As the outputs PRL and PRR therefrom have been given to the controllers 114L and 114R, the velocities VL and VR are controlled to ulti -nately coincide with the target velocities VrL and VrR, respectively The controlling method is described for a mobile robot of power wheeled steering type in the foregoing statement, but it my be similarly applied to those of front-wheel steering, rear-wheel driving type and others.

FIG.8 shows the operation tining and operation nodes of the loconotion nodule. The robot 100 is initialized at tine tl when the power is oi, stays in the initialization node until a predeternined amount of tine TO has elapsed, during which period the current position (x,) is set at the point (0,0), the angle is set as #=0, and necessary control-paraneters and the others are automatically set. After the tine TO has elapsedor after the tine point t2, the feedback control node of the tine Ti and the co-and execution control node of the tine T2 are at the cycle T.The feedback contrl node is assigned to execute the velocity control and position control which have been described in relation to the control system in FIG. 7. The command execution control node is assinged to analyse the content of the conands received from the master via the communication controller 111 and to execute the result of this analysis. Sufficient tine is assigned for all those tasks.The loconotion nodule sequentially repeats the feedback control node and the co-and execution control node at the cycle of the tine T, but, when it receives a reply request signal of a predetermined forn fron the aster ( if a necessity arises), it immediately goes into the comminication control node so as to carry out necessary comminication taslr within a predetermined duration of tine. This communication method is advantageous compared to the one to transnit the state of the locomotion nodule constantly to the connanding section as the communication traffic would not be exessive. The section is returned to the original node after a predetermined tine has passed and kept in that node.

fl.8 shows a case where a communication request signal is sent at tine t3 during the feedback control node and another case where a communication request signal is sent at ti.e t4 during an execution control node. In both cases, the section immediately returned into the previous nodes after the end of the cummunication control rode. By this arrangement, it is apparently possible to execute the travelling operation and the communication operation of the travelling section cocurrently.

FIG.9 is a flow chart to show the communication between the master and the locomotion nodule. The CPtJ 110 of the robot 100 is constantly testing whether a co-and fron the master is conking or not (Step S1). If a cowl cones, each byte of the command is transmitted to the CPU 110 fron the master via the communication controller 111 (Step S2), the step is repeated until the command data has been taken completely (Step S3). The command received is tested whether it is a fast command or not (Step S4). If it is a fast command, it is immediately executed (Step S5). If it is not a fast command, then it should be a slow cord to be stored in a buffer memory BK and registered (Step S6).The CPU 110 tests whether or not the waster bas a request for communication (Step S7). If there is a request, data by one byte, for example the data indicating the current posture of the robot 100, it is sent to the aster byte by byte to conplete the job (Step S8).

FIG. 10 shows a flow chart of the operation at the feedback control node. The CPU 110 constantly is testing whether or not there is a data commu- nication request fron the master (Step S10), and if there is such a request, the CPU 110 reads the pulse data PRL and PRR fron the pulse encoders 117L and 117R to get velocity data (Step S11). The CPU 110 then tests whether the #omotion nodule is in an active state or not (Step S12), and if it is in an active state, the CPU 110 sends the control value VrL and VrR which are calculated fro. the current position x(t), the target position id, etc. (Step S13), and renews the current position ( x(t), y(t), #(t) (Step S14).The CPU 110 further tests whether the locomotion nodule is in an active state or not (Step S15), and if it is not in an active state, the CPU 110 actuates a brake control signal BC to apply brake (Step S16). If the loco option nodule is in an active state, the CPU110 judges whether it is a "GO" command or not (Step S20), and if it is a "GO" command, compares the current position x(t) with the target position xd (Step S21). If it is not a "GO" command, or even if it is a "GO" command the target position xd ( for instance the point E ) is larger than the current position x(t), the CPU 110 tests whether it is a transient state or a stationary state (Step S22) and tests whether the co-and has been completed or not (Step S23). If the command has not been completed, the CPU 110 continues to execute the command and if the command has been completed, the CPU 110 executes a slow command stored in a buffer memory BM (Step S24) and then calculates the next target velocity (Step S25). If the CPU 110 finds the current position x(t) at the Step S21, it is more tban the target position xd, it carries out aforenentioned coordinate trensformtion to complete the job (Step S26).

FIG. 11 is a flow chart to show an example of command execution control node. The CPU 110, whenever step cones to a cc-end control (Step S30), always examines the formet of the command to test whether there is an error or not (Steps S31 and S32), and if there is an error, it recognizes the command as a co-and error (Step S33). If there is not an error, the CPU 110 c#@cks it again (Step S34), and converts the command into the internal format suitable for execution (Step S35), and tests whether it is a slow command or not (Step S36).If it is a slow command, the CPU 110 stores the command in the buffer memory BM (Step 537), and if it is not a slow command but a fast command, the CPU 110 sub-classifies it and executes the co-nd (Step S40).

The types and content of the commands C will now be described below.

A command C is colassified as a slow command SC or a fast command FC. A slow command SC is classified as a "GO" command, a "STOP" command, a "WAIT" command, a "REVERSE" command, "VELOCITY" command, a "CONTROL".

command, a "SET BRAKE" command or a "RESET BRAKE" command. A fast command FC is classified as a "GO 0" command, a "START" command, a "STOP 0" command, a "ADJUST" command, a "VELOCITY 0" command, a "GET" command, a "CANCEL" command, a "FREE MOTOR" command or a "SERVO" command. Each command is describes as follows: (1) "Go" command # Ge, x, y, e; A new coordinate system X'Y' is obtained by parallel-translating the current coordinate system XY and then rotating it by angle 0. The mobile robot 100 is transferred to the new X'-axis and the transient point E thereof is a point (e, 0) or the point where X = e on the old coordinate system.

Effects of various "GO" coins from the locomotion module 10 on, the robot 100 is illustrated is FIGS.12A through 12H.

With the command " GO, 0, 0, 0 ; " which does include any indi- cation for change of direction is given, the robot 100 advances on the travel line X ( =X') as shown in FIG 12A The command " GO, 50, 0, 45 ; " speci the coordinate system X'Y' having the origin at (50, 0) and the revolu tional angle at 450. The robot therefore transfers its course to the axis X' as shownin FIG.12B. As the command " GO, 0, 0, 180 ; " weans the change of direction at the origin (0, 0), the robot 100 travels in the opposite direction as shown in FIG. 120. If the robot 100 receives a "GO" command while travelling, it will .ake a turn as follows.As the command " G60, 100, 0, 30 ; " has the origin at (100, 0) on the new coordinate system X'Y', the inclination at 300 and the transient point E at (60, 0), the robot 100 follows the route indicated in FIG. 120. FIG. 12E illustrates the movement of the robot 100 when the command is for the transient point E at (50, 0) and the inclination angle 8 at 1400. As the command " G110, 100, 0, 90 ; "indicates a transfer at the position E (110, 0) past the new origin (100, 0), the robot 100 loves as shown in FIG. l2F at the rotational angle # = = 900. When the command G100, 0, 40, 0 ; " is received, the robot 100 will love to a new X' -asis fro. the transient point E (100, 0) following the route as shown in FIG.

12G and when a U-turn co-nd M G100, 0, -50, 180 ; " is received, the robot 100 will move as shown is FIG.12H.

(2) "SPIN" command # N # ; This command is effective only when the robot 100 is in a stop state. Otherwise, the co-nd in ignored. It naves the robot spin by the angle 8 and stop after execution of the co-nd. No coordinated transforation is carried out by this command.

(3) "STOP" command # P s ; This command rakes the robot 100 stop at a point (s,0) on the Xaxis. If the robot has already passed the point, it recedes to the alone #int (s,O) and stops thereon. The command is ignored while the robot 100 is in a stop state.

(4) "WAIr command W This command sakes the robot 100 enter into a waiting state.

(5) "REVERSE" command # R e ; With this command, the robot 100 reverses the direction of travelling but no change occurs on the route per se before and after the " reverse " command.

(6) "VELOCITY" command # V v ; This command changes the travelling velocity at the stationary state to determine the velocity of "GO" and "SPIN" cowls which right appear later (7) "CONTRCL" command # C p1, p2 ; This is a slow command to change control parazeters by a servo- loop in order to change the route or required tiie of the travel of the robot 100 at the transient state C1 denotes the type of a parameter and C2 its value.

(8) "SET BRAE " cowl - B; This is the command to apply the brakes 116L and 116R of the robot 100.

(9) "RESET BRAKE" command # K ; This co-nd releases the brakes previously applied on the robot 100.

(10) "STRT" command # S ; This is the commadn to make the robot 100 start travelling by aging the state from waiting to active.

(11) "STOP 0" command # Q This is the co-nd to stop the robot 100 in emergency. Unlike the "Stop" conand, it does not specify any "s (12) "ADJUST" command # Ax, y, # ; This command is similar to the aforementioned "GO" command in transferring onto a new coordinate system. This command is used wihle a "GO" command is being executed, in order to revise the position of the origin or minutely modify the route. This therefore does not specify the transient point E.

(13) "VELOCITY 0" command # U v ; This command changes the velocity of the robot 100 while another command is being executed.

(14) "GET" commadn # T ; This command requests the robot to send back data of the current state. Responding to this command, the robot 100 sends a reply RP.

(15) "CANCEL" command # L ; This is the conand to cancel all slow command SC stored in a buffer memory BM in order not to execute it.

(16) "FREE MOTOR" command # F ; This is the command for moving the robot 100 manually.

(17) "SERVO" command # E ; This is the command for moving the robot 100 with the seno mechanism.

A command sequence shown in FIG. 13 rakes the robot 100 in a stop state at the position P1 advance to the position P2, and recede to the position P3. With the last "START " cowl S, the robot 100 starts travelling.

As is described in the foregoing statements, this invention enables to travel a robot freely with simple commands to realize thereby effective control of the robot. As the travelling section of this invention is separately provided, the similar advantage can be obtained even if commands are sent by radio or similar noans.

It should be understood that many modifications and adaptations of the invention will become apparent to those skilled in the art it is intended to encompass such obvious modifications and chages in the scope of the claiss appended hereto.

Claims (4)

WHAT Is CLAIMED IS:

1. A A locomotion-command method for a mobile robot of the type loving while carrying out communication between a paster and a loconotion nodule which is characterized in that the feedback control tine during which the robot per se executes travelling control and the ccommand execution control tine during which commands froi said aster are alternately assigned for a predeternined tine.

2. The locomotion-command method for a mobile robot as claimed in Claim 1, wherein said co-nd comprises a slow command which is teiporarily stored in a buffer memory and then is executed and a fast command which is immediately executed.

3. The locomotion-command method for a mobile robot as claimed in Claim 1, wherein said locomotion nodule replies to said aster only when a reply request command is sent fron the aster.

4. The locomotiom-command method for a mobile robot as clained in Claim 4, wherein said reply request cord can be sent at an arbitrary tine during the feedback control time of said mobile robot or during the command execution control ode.