CN106094811A - A kind of avoidance obstacle method of assistant robot man-machine interface based on multisensor - Google Patents

Abstract

The invention discloses a kind of avoidance obstacle method of assistant robot man-machine interface based on multisensor；Man-machine interface includes pressure acquisition module, obstacle detection module and controller；Pressure acquisition module is for detecting the active force of operator's hand；Obstacle detection module is used for detecting periphery barrier；Controller is connected with pressure acquisition module and obstacle detection module, for the speed of the data calculating assistant robot that the data gathered according to pressure acquisition module and obstacle detection module detect, and realize walk help and barrier avoiding function according to the speed controlling assistant robot of assistant robot；Avoidance obstacle method includes step: when there is barrier in urgent avoidance district, it is judged that intention power F of operator_IWhether there is the component of X-axis negative direction, if then assistant robot defers to the intention retrogressing avoidance of operator, if otherwise assistant robot will take the mode of lateral translation to evade the barrier in urgent avoidance district.The present invention controls effective, convenient and reliable and with low cost.

Description

A kind of avoidance obstacle method of assistant robot man-machine interface based on multisensor

The applying date of female case: on April 27th, 2015, application number: 2015102098460,

Denomination of invention: a kind of assistant robot man-machine interface based on multisensor and avoidance obstacle method thereof

Technical field

The invention belongs to man-machine interaction, rehabilitation accessory field, more particularly, to a kind of walking aid based on multisensor The avoidance obstacle method of device people's man-machine interface.

Background technology

China, as a populous nation, there are about 1.5 hundred million old peoples, and people with disability 82,960,000, relate to 2.6 hundred million family populations. The crowd of more than 60 years old according to World Health Organization's proposition accounts for society's total number of persons more than 10%, or the crowd of over-65s accounts for society It is senescence country when understanding more than the 7% of total number of persons, according to statistics in 2000, China's coefficient of aged population of more than 60 years old 10.18%, oneself entrance senescence of China is social, will be up to more than 15% to the year two thousand twenty China old people, and the life of old people is good for Kang Wenti, will become social problem.Along with China enters aging society, the old people of walking function obstacle and various senile disease Patient increases year by year, has severely impacted their quality of life, brings white elephant to family and society.

Walking disorder almost puzzlement old people and the serious problems of lower limb disabled orthobiosis, for supporting and Improving the walking ability of various lower limb disorder crowd, walk help lower limb rehabilitation robot has become domestic and international robot and health The study hotspot of multiple medical domain.Lower limb rehabilitation robot has the device such as assistant robot and Intelligent crutch, and ultimate principle is profit Extract the force signal of patient's upper limb with sensing device, judge the intention of patient according to treated force signal, thus help to suffer from Person completes the movement of respective direction.For current sensing device, a kind of scheme is the six-dimensional force sensing utilizing price high Device, improves the cost of assistant robot undoubtedly；Another kind is to form sensor array by multiple one dimension force sensors, passes through The formula of corresponding data and design carries out calculating thus judges the motion intention of people.This scheme can reduce human-machine interface greatly The cost of mouth, but its research is still in the junior stage, and there is various defect in various degree in existing design, it is impossible to effectively The data of power of collection related direction.

Meanwhile, the user of assistant robot system mostly is old people, and their universal vision is poor, and compares ordinary person Operate assistant robot the most mistakenly thus bring potential safety hazard.But current assistant robot system lacks mostly Security mechanism.

Need badly the most at present and develop a set of patient's active force that can effectively detect, cheap, behaviour can be corrected simultaneously The man-machine interface that the issuable mistake of author is intended to is to realize between walk help lower limb rehabilitation robot and patient moving intention Collaborative Control, the adaptive capacity to environment simultaneously strengthened, to obtain more preferable walk help effect, with expect walk help machine is done step-by-step People's commercialization and industrialization.

Summary of the invention

For the defect of prior art, it is an object of the invention to provide effective, the convenient and reliable and low cost of a kind of control The avoidance obstacle method of honest and clean assistant robot man-machine interface.

The invention provides a kind of avoidance obstacle method of assistant robot man-machine interface based on multisensor, described in help The man-machine interface of row robot includes pressure acquisition module, obstacle detection module and controller；Described pressure acquisition module is used for The active force of detection operator's hand；Described obstacle detection module is used for detecting periphery barrier；Described controller is with described Pressure acquisition module and described obstacle detection module connect, for the data gathered according to described pressure acquisition module and institute The data stating the detection of obstacle detection module calculate the speed of assistant robot, and according to the speed controlling of described assistant robot Described assistant robot realizes walk help and barrier avoiding function；It is characterized in that, described avoidance obstacle method comprises the steps: to work as When there is barrier in urgent avoidance district, it is judged that intention power F of operator_IWhether there is the component of X-axis negative direction, the most then help Row robot defers to the intention of operator and retreats avoidance, if it is not, then the mode taking lateral translation is evaded by assistant robot The barrier in urgent avoidance district；Wherein, with assistant robot central point as zero, with the relative dead ahead of assistant robot For X-axis positive direction；With the relative left of assistant robot for Y-axis positive direction.

Further, described urgent avoidance district is rectangle, a length of Length of rectangle, and width is Width, works as laser When distance L between sensor and a sensing point of laser sensor meets following formula, $\left\{\begin{array}{c}\left|L\&times;cos\mathrm{\&alpha;}\right|<Length\\ \left|L\&times;sin\mathrm{\&alpha;}\right|<Width\end{array}\right.;$ Then think Barrier is in urgent avoidance district, and wherein α is certain some angle with the X-axis defined in claim 7 of barrier.

Further, when barrier is in urgent avoidance district, by meeting rectangle condition with right side on the left of statistics X-axis The number of point, few side that selects to count translates；Translation motion detects on translation direction in real time and whether has enough Avoidance space, translation direction or stopping can be changed according to detection case.

Further, described pressure acquisition module includes Force sensor, Signal-regulated kinase and data acquisition module； Described Force sensor is two layers of tube-in-tube structure, and outer layer is the internal cylinder being become cuboid by hollow out, and internal layer is square column Body, has gap between cylinder and square cylinder, be respectively arranged with 1 FSR sensor in four sides of internal layer square cylinder Fill described gap；Component for acquisition operations person's right-hand man's active force vertical and horizontal；Described Signal-regulated kinase is used for The active force gathering described Force sensor exports after nursing one's health；Described data acquisition module is for after conditioning Data obtain magnitude of voltage after being AD converted.

Further, described obstacle detection module includes laser sensor and for connecting described laser sensor USB high speed data lines with described controller；Described laser sensor is for gathering the obstacle information of surrounding enviroment, and passes through The obstacle information of collection is sent to controller by described USB data line.

Further, described controller is for barrier letter obstacle detection module gathered according to Artificial Potential Field Method Breath is converted to repulsion；Calculating repulsion is made a concerted effort with operator's active force, and will make a concerted effort to be converted to for controlling assistant robot in fact The speed of existing walk help.

It is an advantage of the current invention that and first utilize cheap FSR sensor and design in operator's active force context of detection Force sensor instead of traditional six-axis force sensor, can effectively detect again the effect of operator while reducing cost Power；Utilizing laser sensor to be scanned periphery barrier, operator is drawn by the obstacle information using scanning to obtain Lead and correct, it is ensured that operator and the safety of assistant robot；The control method utilizing multithreading to realize meets simultaneously and helps Row robot obstacle-avoiding and the demand of walk help, and ensure that real-time；Relatively simple structure makes this assistant robot man-machine Interface goes for the assistant robot of various frame for movement, it is easy to popularization and application, meets the demand that user is different.

Accompanying drawing explanation

Fig. 1 is the multisensor assistant robot man-machine interface system block diagram that the present invention provides.

Fig. 2 is an application example schematic diagram of the embodiment of the present invention.

Fig. 3 is FSR used in the present invention (Force Sensing Registers, power sensing record unit) sensor Machinery schematic diagram.

Fig. 4 is the schematic diagram of the signal conditioning circuit of FSR sensor used in the present invention, and wherein (a) is amplifying circuit Physical circuit figure, (b) is the physical circuit figure of generating circuit from reference voltage.

Fig. 5 is the force analysis figure that the embodiment of the present invention is applied to an assistant robot example.

Fig. 6 is the fundamental diagram of the laser sensor that the embodiment of the present invention is used.

Fig. 7 is the structural representation of barrier described in the invention.

Fig. 8 is the schematic diagram that the barrier that the present invention is applied merges algorithm；Wherein (a) is that two adjacent barrier merge Front method for expressing, (b) is the method for expressing after two adjacent barrier merge.

Fig. 9 is the urgent obstacle-avoidance area schematic diagram that the urgent avoidance scheme that the present invention proposes is arranged.

Figure 10 is the state transition graph of the urgent avoidance scheme that the present invention proposes.

Figure 11 is the multithread programs flow chart of the embodiment of the present invention.

Detailed description of the invention

In order to make the purpose of the present invention, technical scheme and advantage clearer, below in conjunction with drawings and Examples, right The present invention is further elaborated.Should be appreciated that specific embodiment described herein only in order to explain the present invention, and It is not used in the restriction present invention.

The invention provides a kind of avoidance controlling assistant robot man-machine interface effective, convenient and reliable and with low cost Control method.

As it is shown in figure 1, man-machine interface includes pressure acquisition module 1, obstacle detection module 2 and controller 3；Pressure is adopted Collection module 1 is for detecting the active force of operator's hand；Obstacle detection module 2 is used for detecting periphery barrier；Controller 3 with Pressure acquisition module 1 and obstacle detection module 2 connect, for the data gathered according to pressure acquisition module 1 and barrier The data of detecting module 2 detection calculate the speed of assistant robot, and according to walking aid described in the speed controlling of assistant robot Device people realizes walk help and barrier avoiding function.

Pressure acquisition module 1 includes Force sensor 11, Signal-regulated kinase 12 and the data acquisition module being sequentially connected with 13；Force sensor 11 is two layers of tube-in-tube structure, and outer layer is the internal cylinder being become cuboid by hollow out, and internal layer is square column Body；There is gap between cylinder and square cylinder, be respectively arranged with 1 FSR in four sides of the square cylinder of each layer and pass Sensor fills described gap；Force sensor 11 is for the component of acquisition operations person's right-hand man's active force vertical and horizontal；Signal Conditioning module 12 exports after the active force gathering described Force sensor is nursed one's health；Data acquisition module 13 is for right Data after conditioning obtain magnitude of voltage after being AD converted.

Obstacle detection module 2 includes laser sensor 21 and for connecting described laser sensor 21 and controller 3 High speed USB data line；Laser sensor 21 is for gathering the obstacle information of surrounding enviroment, and passes through described high speed USB number According to line, the obstacle information of collection is sent to controller.

Controller 3 is for being converted to repulsion according to Artificial Potential Field Method by the obstacle information that obstacle detection module 2 gathers； Calculating repulsion is made a concerted effort with operator's active force, and is converted to realize the speed of walk help for controlling assistant robot by making a concerted effort.

Fig. 2 shows an application example of the assistant robot man-machine interface that the embodiment of the present invention provides；For pressure Acquisition module 1, Force sensor 11 includes that 8 FSR ((Force Sensing Registers, power sensing record unit)) pass Sensor, its frame for movement is as it is shown on figure 3, be the Force sensor through specified arrangement and machining；Binding signal conditioning mould Block 12 and data acquisition module 13, pressure acquisition module 1 can detect the active force of operator's hand.Obstacle detection module 2 It is made up of laser sensor 21 and the high speed USB data line for being connected with controller 3, periphery barrier can be detected；Control Industrial computer selected by device 3, and it communicates with pressure acquisition module 1 and obstacle detection 2 module, and the data gathering them are carried out Comprehensive analysis and calculating, obtain a suitable speed and be sent to external communication interface, and can draw the operation of operator Lead, it is ensured that control safety.The external communication interface being connected with controller is used for being connected omnidirectional's motor chassis in the present embodiment, electricity Machine chassis can drive the motion of assistant robot.

Fig. 3 is the mechanical schematic diagram of the Force sensor 11 in pressure acquisition module 1 of the present invention, is demonstrated by Fig. 2 operation Person holds part, i.e. the detailed design of Force sensor 11.Force sensor 11 is the two layers of tube-in-tube structure in two, left and right, and outer layer is The internal cylinder being become cuboid by hollow out, internal layer is square cylinder；The FSR sensor that 4 processes are uniformly processed is glued respectively It is attached to before and after an internal layer rectangle cylinder on four sides, left and right.Certain gap is had, this seam between internal layer and outer layer Gap can be filled just by 4 sensors processed being attached to square cylinder side.In embodiments of the present invention, FSR sensing The tow sides of device have pasted the rubber of flexible smooth, and the outside of rubber is part of the force, and the FSR sensor being processed as is permissible Fully external pressure is reacted in the change of resistance, thus produces the obvious voltage signal of change.When the hands cylinder to outside When body applies the thrust of any direction, Force decomposition can be longitudinally (front and back) and horizontal by this Force sensor 11 The component of (left and right).Force sensor 11 is made of metal, and therefore it will not produce deformation because of grip, has well filtered power Grip useless in signal, and obtain thrust or the pulling force needed.

Fig. 4 is the schematic diagram of the Signal-regulated kinase in the present invention.Signal-regulated kinase includes simple amplifying circuit and base Quasi-voltage generation circuit.Reference voltage generating circuit is made up of bleeder circuit and voltage follower circuit, and its outfan is access in Positive input V as the reference voltage to simple amplifying circuit_ref.Simple amplifying circuit be by operational amplifier form same Phase amplifying circuit, FSR sensor is linked in in-phase amplification circuit as resistance, if its resistance is R_fsr.Consult FSR sensing The product information of device understands the size of pressure and resistance and is approximated to inverse proportion functional relationship, along with the increase of pressure, FSR sensor Resistance will reduce rapidly.If F=N × R_fsr ^-1……(1)；In-phase amplifier output voltage V_oIt is represented by: V_o=V_ref×(1 +R₁×R_fsr ^-1)……(2)；Formula (1) is substituted into formula (2) can obtain: V_o=V_ref×(1+R₁×F×N^-1)……(3)；Wherein F is Act on the pressure size on FSR sensor；N is proportionality coefficient；R₁Variable-resistance resistance for amplifying circuit simple in Fig. 4 Value.

From formula (1) (2), when pressure is the least, the resistance R of the FSR sensor in in-phase amplification circuit_fsrIt is much larger than Variable resistance R₁Value, the output of voltage follower is reference voltage V_ref.By formula (3) it can be seen that the present invention use electricity Road design can allow pressure become linear relationship with the relation of output voltage within the specific limits, and output voltage is along with the increasing of pressure Increase greatly and linearly.The input interface of circuit is divided into two groups, often organizes and all can be accessed by 4 FSR sensor (totally 8), just can connect Enter necessary 8 FSR sensor of Force sensor 11.Output interface has 10 pins, including the output of No. 8 voltage followers Voltage and two ground wires.Output interface is connected with data acquisition module 13 (as a example by the USB4716 grinding China), therefore output pin The arrangement of the analog voltage input pin put in order with data acquisition module 13 consistent so that wiring.

Data acquisition module 13 used in the present invention can Shi Yanhua company produce data collecting card USB4716, can To change any one data collecting card that can carry out the AD conversion more than 12 as required into.Due to data collecting card USB4716 is the most ripe product, and therefore without any change on hardware, what needs did is exactly according to product description Writing control program allows it complete the work of corresponding AD conversion.

Shown in Fig. 5 is the force analysis figure of embodiment of the present invention pressure acquisition module 11, as recited in claim 7, Forward for X-axis positive direction, to be Y-axis positive direction to the left, set up the local coordinate system of assistant robot, pressure acquisition module and barrier Hinder thing detecting module consistent with the coordinate system on the chassis used in the embodiment of the present invention, because unified coordinate system is conducive to letter Change and calculate.In conjunction with Fig. 2 it can be readily appreciated that in Fig. 5 FHL and FHR represent the active force of operator's left hand and right hand, wherein F respectively_lX,F_lY, F_rX,F_rYIt is that FHL and FHR is along X-axis and the decomposition of Y direction respectively.F_RIt it is the barrier detected according to obstacle detection module The repulsion that assistant robot is produced by the preceding object thing that information is calculated.The unified coordinate system set up according to the present embodiment, In conjunction with Fig. 5, if it is F that left side sleeve is positioned at the force value measured by the FSR sensor in X-axis positive direction₁, respectively The measured value of individual FSR sensor is followed successively by F₁～F₄, in like manner, the measured value of the right sleeve FSR sensor is F₅～F₈.Thus I Can be by the active force of operator right-hand man along the component F of X-axis and Y-axis₁～F₈Numerical value plus-minus represent, such as formula (4) institute Show: $\left\{\begin{array}{c}{F}_{1X}=F_3-F_1\\ F_{1Y}=F_4-F_2\\ F_{rX}=F_7-F_5\\ F_{rY}=F_8-F_6\end{array}\right.$ (4)。

The movement of assistant robot can be divided mainly into anterior-posterior translation, left and right translation and rotary motion.Force sensor gathers Data be that the power of two handss is respectively at X-axis and the component of Y-axis active force.The repulsion that barrier produces is carried out point by controller 3 Solve and calculate and make a concerted effort, then can control the motion of assistant robot according to following basic skills:

(1) if X-direction make a concerted effort be not 0 and right-hand man's force direction consistent, then assistant robot along X-axis make a concerted effort side To motion；

(2) if Y direction makes a concerted effort not to be 0, then assistant robot moves along the resultant direction of Y-axis；

(3) if right-hand man's force direction of X-axis is contrary, then calculating moment according to the arm of force, assistant robot is along moment side To rotating；

Above three basic skills can in any combination, so that assistant robot completes all of base on two dimensional surface This motion.

Laser sensor 21 in the embodiment of the present invention be Japan Bei Yang company produce model be URG-04LX-UG01 Laser sensor, as shown in Figure 6, by the unified coordinate system set up before, definition X-axis positive direction is 0 ° to its fundamental diagram, that This laser sensor can realize carrying out the comprehensive scanning from-120 ° to 120 ° counterclockwise.Laser sensor 21 Effect be the barrier of periphery of scanning assistant robot, and calculate the repulsion of barrier according to Artificial Potential Field Method, the most such as Under:

The fundamental formular that Artificial Potential Field Method repulsion calculates is as follows: F_ri=K × (r-R₀)^-n……(5)；K, R in formula₀, n be normal Number, typically takes positive integer.R is the distance of barrier and robot, F_riSize for repulsion.

For each barrier in environment, design six tuple structure rEdge, rAngle, lEdge, lAngle, Force, Angle} are used as the location contents in barrier chained list.As it is shown in fig. 7, laser sensor 21 is at O point, rEdge table Show and represent on the right of barrier along with laser sensor 21 even along distance | the OB |, rAngle with laser sensor 21 on the right of barrier Line and the angle ∠ BOX of X-axis positive direction.In like manner, lEdge Yu lAngle represent respectively distance value | OA | of barrier left margin with Angle ∠ AOX.Force represents the size of repulsion, and Angle represents the direction of repulsion and the angle of X-axis positive axis.Herein for Each barrier, chooses the ∠ AOB angular bisector OD direction of barrier edge and laser sensor 21 composition as repulsion direction, R=| OD | substitutes into formula (5), the size of the repulsion Force that available single barrier produces: F_ri=K × (| OD |-R₀)^-n…… (6)；It will be noted that when two obstacle distances are close, the gap between them may be not enough to allow assistant robot from Middle safety.The repulsion produced due to barrier is only relevant with the relative distance of it and assistant robot, the barrier in same place The repulsion hindering thing group to produce may be far longer than the active force of operator after vector is cumulative, thus hinders walk help machine People normally advances.For avoiding the double counting of the repulsion in close direction, it should barrier close together is merged process. Fig. 8 gives the example that a barrier merges.

For two adjacent barrier such as Fig. 8 (a), six tuple structure according to designing before are readily obtained ∠ BOC's Value α, according to the cosine law, obtains gap | BC | of barrier 1 and barrier 2:

$\left|BC\right|=\sqrt{{\left|OB\right|}^2+{\left|OC\right|}^2-2\left|OB\right|\left|OC\right|cos\mathrm{\&alpha;}}\mathrm{......}\left(7\right);$

If | BC | is less than the size of the assistant robot of application the present embodiment, then barrier is merged, such as Fig. 8 Shown in (b).Using the right of barrier in Fig. 8 (a) 1 along the left margin OA of OD and barrier 2 as the edge merging obstruction, OE is the angular bisector of △ AOD, can obtain OE length | OE | according to the angular bisector formula of triangle, and bringing (5) into can obtain To merging the repulsion F that obstruction produces_ri=K × (| OE |-R₀)^-n……(8)；

Note, calculate | OE | rather than directly take the scan values of OE direction laser sensor 21 and be because line segment OE extended line May be just passed through the gap of two barriers, now laser sensor 21 can't detect barrier, may affect repulsion Calculate.

Barrier is ultimately stored in obstacle information chained list after merging, final repulsion F_rProduced by each barrier Raw repulsion F_obiVector sum:I is the i-th barrier in obstacle information chained list, and n is The sum of barrier；

In practice, for the ease of calculating and merging with operator's hand active force, we calculate repulsion respectively Component in X-axis and Y direction is also sued for peace.The direction definition of the positive and negative and coordinate axes of power is consistent. $F_{obX}=\underset{i=0}{\overset{n}{\&Sigma;}}{F}_{obiX}=-\underset{i=0}{\overset{n}{\&Sigma;}}\left|F_{obi}\right|{\mathrm{cos\&theta;}}_i\mathrm{......}\left(10\right),$ $F_{obY}=\underset{i=0}{\overset{n}{\&Sigma;}}{F}_{obiY}=-\underset{i=0}{\overset{n}{\&Sigma;}}\left|F_{obi}\right|{\mathrm{sin\&theta;}}_i\mathrm{......}$ $\left(11\right);$ Final To repulsion vector be F_r=[F_obX F_obY 0]^T。

In description to Fig. 5 before, we have been obtained for representing the variable F of operator's both hands force component_lX, F_lY,F_rX,F_rY.We can be with four dimensional vector F_SRepresent the applied force vector input as man-machine interface of operator Amount, by three-dimensional vector F_IRepresent the intention of operator, claim vector F_IIntention force vector for operator.Then have: F_s=[F_lX F_lY F_rX F_rY]^T……(12)；F_I=[F_X F_Y M_Z]^T……(13)；

For F_I, F_XRepresent operator's intention power in X-direction, F_YRepresent operator's intention power in Y direction, M_Z Represent the rotation torque that assistant robot is applied by operator, be just in a counterclockwise direction.According to Artificial Potential Field Method before in conjunction with The repulsion vector F calculated_r, have equation below to set up: F_I=E_ISF_S+F_r……(14)；

The applied force vector of operator and the repulsion vector calculated of Artificial Potential Field Method are combined by formula (14), as The intention force vector of operator.Wherein E_ISIt is the transition matrix of 3 × 4: $E_{IS}=\left[\begin{array}{cccc}1& 0& 1& 0\\ 0& 1& 0& 1\\ -L& 0& L& 0\end{array}\right]\mathrm{......}\left(15\right),$ Wherein L Represent the half of assistant robot width shown in Fig. 5.

Obtain operator and be intended to force vector F_IAfter, use open loop admittance control algorithms to be changed by the intention force vector of operator For velocity vector.If the intention force vector of operator is F_I(s), the velocity vector of assistant robot output is V_(s), then according to leading The transmission function of control algolithm received can obtain:Wherein K is proportionality coefficient, and τ is time constant.

Write as continued time domain model to have: $\mathrm{\&tau;}\&CenterDot;\frac{dV\left(t\right)}{dt}+V\left(t\right)=K\&CenterDot;F_I\left(t\right)\mathrm{......}\left(17\right);$

Discretization: $\mathrm{\&tau;}\&CenterDot;\frac{V\left(n\right)-V(n-1)}{T}+V\left(n\right)=K\&CenterDot;F_I\left(n\right)\mathrm{......}\left(18\right);$ Abbreviation:

$V\left(n\right)=K\&CenterDot;\frac{T}{T+\mathrm{\&tau;}}\&CenterDot;F_I\left(n\right)+\frac{\mathrm{\&tau;}}{T+\mathrm{\&tau;}}\&CenterDot;V(n-1)\mathrm{......}\left(19\right);$

Artificial Potential Field Method combines with admittance control algorithms and is applied to assistant robot distance barrier situation farther out, can Effectively to guide operator to keep at a distance with periphery barrier, simultaneously in order to ensure the safety of assistant robot man-machine interface, The present embodiment devises a kind of urgent avoidance scheme, and details are as follows:

When barrier and assistant robot hypotelorism to such an extent as at any time it may happen that during collision, it is necessary to take urgent avoidance Measure is to ensure operator and the safety of assistant robot.First the present embodiment divides urgent avoidance district, when in urgent avoidance district When finding barrier, if the intention force vector F of operator_IHaving the component of X-axis negative direction, assistant robot will comply with operator Intention retreat avoidance, otherwise the mode taking lateral translation is attempted evading the obstacle in urgent avoidance district by assistant robot Thing.

The choosing method in urgent avoidance district uses rectangle condition, as it is shown in figure 9, urgent obstacle-avoidance area is the rectangle in Fig. 9, Laser sensor 21 is at O point.Rectangle is then judged when laser sensor 21 exists when any one scanning element meets rectangle condition There is barrier in region, rectangle condition is described as follows:

If (x y) is a sensing point of laser sensor, | OA |=L to A.If rectangle length is Length, width is Width, when the part that cooks noodle is set up, rectangle condition is set up； $\left\{\begin{array}{c}\left|L\&times;cos\mathrm{\&alpha;}\right|<Length\\ \left|L\&times;sin\mathrm{\&alpha;}\right|<Width\end{array}\right.\mathrm{...}\left(20\right),$ α is certain point of barrier Angle with X-axis.

Now start urgent avoidance scheme.First urgent avoidance scheme selects the positive direction of Y-axis or negative direction to translate Avoidance, set direction make use of a kind of Voting Algorithm, and i.e. with X-axis as boundary, on the left of statistics X-axis and right side meets the point of rectangle condition Number, select few side of counting to translate.The most this Voting Algorithm have selected one of assistant robot initially State.During translation avoidance, assistant robot detects assistant robot distance such as Fig. 9 to edge body of wall, once in real time Find that this distance, less than safe distance, changes kinestate the most immediately.Obstacle is there is in assistant robot at urgent obstacle-avoidance area Can change between various states during thing, concrete urgent avoidance scheme can be with State Transferring chart as shown in Figure 10 Showing, table below is described in detail the various conditions of various state and State Transferring.

State	Explanation
		0	Clear in urgent avoidance district, performs admittance control algorithms and Artificial Potential Field Method
1	Urgent avoidance district has barrier, robot to move right
		2	Urgent avoidance district has barrier, robot to be moved to the left
3	Assistant robot emergent stopping

The embodiment of the present invention uses multithreading to realize above scheme, and what method described above can be brief is summarised as Multithread programs flow chart shown in Figure 11.After controller initializes, main thread and a thread successively start, and they are independent, and Row work, and all carry out periodic duty.Two threads use repulsion vector F_rIt is right jointly to complete with state variable State The control of assistant robot.State variable State initial value is 0, represents that transport condition is normal.

The job step propping up thread is as follows:

First (result is an array set, each element representation in set to extract the scanning result of laser sensor The obstacle distance of certain angle in sweep limits), according to scanning result, in first determining whether urgent avoidance district, whether there is obstacle Thing, if it is present, change the value of state variable State according to the urgent obstacle avoidance algorithm described in Figure 10.

If there is no, then according in Fig. 7 set up barrier model and barrier six tuple structure rEdge, RAngle, lEdge, lAngle, Force, Angle}, first can obtain rEdge=| OB |, rAngle=from laser sensor ∠ BOX, lEdge=| OA |, lAngle=∠ AOX, according to rAngle and lAngle can obtain Angle=∠ DOX=1/2 × (∠ AOX+ ∠ BOX), and then distance value | OD | corresponding to Angle, recycling can be read according to Angle by laser sensor (6) formula calculates the size of repulsion Force.The information of single pass all of periphery obstacle can be calculated by the above process, formed Barrier chained list, the location contents of chained list is above-mentioned six tuple structure.

After setting up barrier chained list, the cosine law (formula (7)) is utilized to calculate the gap between adjacent barrier (in Fig. 8 (a) BC) assistant robot whether can be made to pass through, if could not; two barriers would be merged into one and carry out processing (Fig. 8 (b)), and recalculate repulsion according to formula (8).After barrier chained list is arranged, all barriers in chained list are produced Repulsion carries out vector summation and decomposes (formula (9) (10) (11)) along the coordinate axes set, and obtains all obstacles in barrier chained list The vector F that makes a concerted effort of the repulsion that produce are raw_r=[F_obX F_obY 0]^T, finally change the value of state variable State into 0.It it is more than branch line The job content in one cycle of journey.

Main thread job step is as follows:

After main thread starts AD conversion, the 8 road FSR sensings that AD conversion result, i.e. data acquisition module 13 gather are read in timing The active force numerical value F that device detects₁～F₈, utilize formula (4) to calculate；F_s=[F_lX F_lY F_rX F_rY]^T, according to the repulsion of a thread Vector result of calculation F_r=[F_obX F_obY 0]^TWith the intention force vector F that formula (14) calculates operator_I=[F_X F_Y M_Z]^T, then root Final output speed V is calculated according to formula (19)_N=[V_X V_Y V_Z]^T.The value of last reading state variable State, will if 0 Result of calculation V_NBeing sent to base plate electric machine, assistant robot is with speed V_NMotion；If 1, assistant robot with fixed speed to Right translation avoidance；If 2, assistant robot with fixed speed to left avoidance；If 3, assistant robot promptly stops Only.

In sum, this detailed description of the invention have simple in construction, easy-to-use, versatility is high, real-time and safety good, and The feature that cost is the cheapest.Can be applied to easily in general assistant robot system thus strengthen locomotor activity damage and The locomotor activity of amblyopia patient, improves their quality of life.

As it will be easily appreciated by one skilled in the art that and the foregoing is only presently preferred embodiments of the present invention, not in order to Limit the present invention, all any amendment, equivalent and improvement etc. made within the spirit and principles in the present invention, all should comprise Within protection scope of the present invention.

Claims (6)

1. an avoidance obstacle method for assistant robot man-machine interface based on multisensor, described assistant robot human-machine interface Mouth includes pressure acquisition module (1), obstacle detection module (2) and controller (3)；Described pressure acquisition module (1) is used for examining Survey the active force of operator's hand；Described obstacle detection module (2) is used for detecting periphery barrier；Described controller (3) with Described pressure acquisition module (1) and described obstacle detection module (2) connect, for adopting according to described pressure acquisition module (1) The data that the data of collection and described obstacle detection module (2) detect calculate the speed of assistant robot, and help according to described Assistant robot described in the speed controlling of row robot realizes walk help and barrier avoiding function；It is characterized in that, described avoidance obstacle side Method comprises the steps:

When there is barrier in urgent avoidance district, it is judged that intention power F of operator_IWhether there is the component of X-axis negative direction, if It is that then assistant robot defers to the intention retrogressing avoidance of operator, if it is not, then assistant robot will take the mode of lateral translation Evade the barrier in urgent avoidance district；

Wherein, with assistant robot central point as zero, with the relative dead ahead of assistant robot for X-axis positive direction；With The relative left of assistant robot is Y-axis positive direction.

2. avoidance obstacle method as claimed in claim 1, it is characterised in that described urgent avoidance district is rectangle, the length of rectangle Degree is Length, and width is Width, under distance L between laser sensor and a sensing point of laser sensor meets During formula, $\left\{\begin{array}{c}|L\&times;\cos \mathrm{\&alpha;}|<\mathrm{Lengt}\\ |L\&times;\sin \mathrm{\&alpha;}|<\mathrm{width}\end{array};\right.$ Then think barrier in urgent avoidance district, wherein α be barrier certain point with right want Seek the angle of X-axis defined in 7.

3. avoidance obstacle method as claimed in claim 1 or 2, it is characterised in that when barrier is in urgent avoidance district, logical Crossing on the left of statistics X-axis and right side meets the number of point of rectangle condition, few side that selects to count translates；At translation motion In in real time whether have enough avoidance spaces on detection translation direction, translation direction or stopping can be changed according to detection case.

4. assistant robot man-machine interface as claimed in claim 1, it is characterised in that described pressure acquisition module (1) includes Force sensor (11), Signal-regulated kinase (12) and data acquisition module (13)；

Described Force sensor (11) is two layers of tube-in-tube structure, and outer layer is the internal cylinder being become cuboid by hollow out, internal layer It is square cylinder, between cylinder and square cylinder, has gap, be respectively arranged with 1 in four sides of internal layer square cylinder FSR sensor fills described gap；Component for acquisition operations person's right-hand man's active force vertical and horizontal；

Described Signal-regulated kinase (12) exports after the active force gathering described Force sensor (11) is nursed one's health；

Described data acquisition module (13) obtains magnitude of voltage after being AD converted the data after conditioning.

5. assistant robot man-machine interface as claimed in claim 1, it is characterised in that described obstacle detection module (2) is wrapped Include laser sensor (21) and be used for the USB high speed data lines connecting described laser sensor (21) with described controller (3)； Described laser sensor (21) is for gathering the obstacle information of surrounding enviroment, and the barrier that will be gathered by described USB data line Thing information is hindered to be sent to controller (3).

6. assistant robot man-machine interface as claimed in claim 1, it is characterised in that described controller is for according to artificial gesture The obstacle information that obstacle detection module (2) gathers is converted to repulsion by field method；Calculate the conjunction of repulsion and operator's active force Power, and be converted to realize the speed of walk help for controlling assistant robot by making a concerted effort.