Dispersing robots in an unknown environment

Dispersing robots in an unknown environment

Ryan Morlok and Maria Gini

Department of Computer Science and Engineering,University of Minnesota,200 Union St.S.E.,Minneapolis,MN55455-0159{morlok,gini}@https://www.wendangku.net/doc/e516576440.html, Summary.We examine how the choice of the movement algorithm can a?ect the success of a swarm of simple mobile robots attempting to disperse themselves in an unknown environment.We assume there is no central control,and the robots have limited processing power,simple sensors,and no active communication.We evaluate di?erent movement algorithms based on the percentage of the environment that the group of robots succeeds in observing.

1Introduction

The problem we address is that of dispersing a group of mobile robots in an unknown environment.We assume the robots do not know how many other robots are operating in the same environment,where those robots are located, and where those robots have been.

The primary motivation for this work comes from the Scout project[9]. The scouts are small,two wheeled robots with extremely limited processing capability.In general,the scouts are deployed by being hauled or launched into the environment by a larger robot.Their job is then to disperse throughout the environment so that it can be e?ectively monitored.Currently the scouts are teleoperated,but they can also perform some autonomous operations,such as hiding and watching for motion[9],by proxy processing over a radio link.

One of the major issues with these type of robots is communication.Since the robots are small,and therefore do not have a great deal of available elec-trical power,it can be di?cult(and in some cases impossible)to generate a signal strong enough to communicate with all other robots.This problem is worsened by the fact that the scouts are designed to explore hostile environ-ments,which may have physical characteristics that further hamper any sort of radio based communication.Because of this,we will constrain our algo-rithms not to require any explicit communication,and to use the sensors to communicate implicitly by observing cues from the environment.This type of communication,which is called stigmergy in the biology literature,is common

2Ryan Morlok and Maria Gini

in swarm approaches to robotics[1].We will further assume that the robots have enough local processing power,so all computation is done locally.

2Related Work

Coverage of terrain during motion is important in many application domains, such as?oor cleaning,lawn mowing,harvesting,etc.A recent survey[2]clas-si?es the existing algorithms for terrain coverage.

Wagner et al.[11]formalize the terrain covering problem and propose two algorithms.one called mark and cover(MAC),the second called probabilistic coverage(PC),both for single and multiple robots.They show how several cooperating robots can obtain faster coverage.Algorithms inspired by insect behaviors,such as ants,are becoming popular both for terrain coverage[6], where robots leave trails and cover the terrain repeatedly,and for optimization of paths[8].

The study by Hsiang et al.[4,5]is the closest to our work.In their work, they examined methods for dispersing robots from?xed locations to cover the entire environment.They assume a continuous stream of robots would be en-tering the environment through speci?c,predetermined locations.The goal of the robots would then be to position themselves such that the entire area of the accessible space is covered.While this work has great properties/guarantees, it is not immediately applicable to the problem we are investigating in this paper.The reason is that while each robot only has extremely local knowl-edge of the environment,through the use of the deterministic movement and in?nite supply of robots the information available at the point at which any given robot is located is su?cient to guarantee that the robot will make the correct choice.In our investigation,it would be impractical to assume that there are enough robots to cover the entire map and to guarantee that every robot can remain within sensor range of the other robots.In our experiments, we assume there are at most50robots present in the environment.In envi-ronments such as the Hospital World(shown later in Figure2),this would allow possibly two robots per room explored.Clearly there are not enough robots to make Hsiang’s algorithm feasible.

3Motion algorithms

The purpose of this study is to examine how the selection of the movement algorithm for a multi-robot system a?ects the coverage of robot observation in a variety of environments.

We considered four distinct movement algorithms,all of them reactive in nature.Each movement algorithm controls two types of movements:for-ward/backward and turning.Turning can occur in place or while the robot is moving.The sensors available to the movement algorithms are16sonar range

Dispersing robots in an unknown environment 3

Fig.1.Left:A robot using the FollowWall algorithm navigates a corner.Center:A robot using the SeekOpen algorithm calculates the average obstacle vector and moves in the direction opposite from this vector.Right:The Fiducial movement algorithm.(a)A robot detects another robot behind it within its sensor range,and begins to adjust its forward motion to turn so the detected robot will be immediately behind.(b)The detector robot has successfully positioned the detected robot behind it;the detector robot will now continue with straight forward movement.

?nders,each of which returns the distance of the nearest object detected in the direction in which the range ?nder is pointing.A ?ducial range ?nder is also used to determine the location of other robots,which is needed for the Fiducial algorithm.Robots have only local knowledge,they are not under global control (no central source knows the state of all of the robots),and do not have any knowledge of the environment other than what they can detect with their sensors.

3.1Random Walk

The RandomWalk movement algorithm is the most basic of the algorithms we examined.A robot using this algorithm can be in one of two states:random forward movement or obstacle avoidance.In random forward movement,the robot moves forward with a small random turn factor between ?10?and 10?(the robot’s path is curved)which is changed at random intervals,ranging between 10s and 15s.The amount of the turn is constrained to ensure the robot does not end up going in small circles.Once the robot detects that it has encountered an obstacle,it enters the obstacle avoidance state.In this state,the robot will stop,turn a random amount (in the range 120?to 240?),and transition back to the standard forward movement state.

3.2Follow Wall

The idea behind the FollowWall algorithm comes from the fact that in many indoor environments,if a robot could ?nd an outer wall of the building and follow it,the robot would be led through much of the structure.A robot

4Ryan Morlok and Maria Gini

using the FollowWall algorithm will search for an obstacle(presumably a wall or large object in the environment)and then proceed to follow that wall inde?nitely.In this algorithm,a robot has four states:?nd wall,align to wall, follow wall,and navigate corner.If the robot believes that it has lost the wall in any of the three non-?nd-wall states,it will reset back to the initial?nd wall state and search for a new wall to follow.

The major problem of this movement algorithm is that it assumes every obstacle encountered is a wall,rather than trying to determine if the observed entity is something smaller,such as a robot at close range.Because of this, when many robots using this algorithm are together,they will tend to perceive each other as walls and try to align themselves to each other.This is wasteful, since the alignment procedure will not e?ectively spread the robots out in the environment.

3.3Seek Open

The SeekOpen movement algorithm causes a robot to move toward open areas in the map.The motivation for the SeekOpen algorithm is similar to the?ducial robot avoidance algorithm(discussed next).According to the assumptions of the experiment,all robots start out in the same general area, grouped fairly close together.Because of this,all the robots tend to have objects(generally other robots)close to themselves at the beginning of a run. The goal of the seek open algorithm is to motivate the robots to disperse as quickly as possible.

SeekOpen is executed by?rst calculating the average obstacle vector for all obstacles in sensor range.The average obstacle vector is computed by summing the vectors pointing to all of the objects within sensor range and dividing by the number of vectors summed.The magnitude of the vector must be large for objects close to the robot and small for objects far away.This is accomplished by setting the magnitude of a perceived obstacle vector equal to the maximum range of the sensors minus the perceived distance a given obstacle is from the robot,or by using some other function which decreases with distance,as done when using arti?cial potential?elds for navigation[7]. After the average obstacle vector is computed,the goal of the robot becomes to move in the opposite direction of the average obstacle vector.The robot turns toward the direction of the negative obstacle vector.The rate of turn is determined by the magnitude of the average obstacle vector.This allows the SeekOpen algorithm to not run into walls as well as disperse from other robots.This is illustrated in Figure1.

3.4Fiducial

The Fiducial movement algorithm was inspired by the idea that the robots would be able to recognize other robots,and therefore move away from them. The original concept involved a simple signal(possibly a weak radio signal)

Dispersing robots in an unknown environment5 that each robot would emit,so that another robot could detect the signal,and determine an approximate distance to the originating robot based solely on signal intensity.The problem of moving away from other robots would then become a goal of?nding areas in which signal intensities are low,which can be done by any hill-climbing algorithm.Unfortunately,there was no straight-forward way to implement such as system within Stage,and therefore an alternative method was sought.

The solution to the simulation problem was to use a?ducial device(gen-erally used to?nd beacons in the Stage simulator)and attach a beacon to every robot.This allows a given robot to know the polar coordinates of other robots within sensor range(sensor range is a semi-circle of?xed radius)with respect to its own position and orientation.The information can be used to steer away from other robots.

With the?ducial information,implementation of an avoidance algorithm is straightforward.Whenever a robot detects another robot within sensor range, the robot adjusts its movement so that it is moving away from the detected robot.When no robots are in sensor range,a robot simply moves according to the random walk algorithm.If at any time a robot encounters a physical obstacle such as a wall,the obstacle avoidance technique takes precedence over whatever movement algorithm the robot is currently executing.

4Simulation Environment and Data Analysis

To compare the algorithms,we performed a large number of experiments within the Player/Stage[10,3]simulator.The virtual robot used for experi-ments is a rectangular,four-wheeled,Pioneer-like robot.Although the moti-vation for this work comes from the scout project,the scouts are not currently modeled in Stage,and this motivated the change in platform.The robots used in the simulations have16sonar range?nder sensors,a pan-tilt-zoom camera with blob?nding software capabilities(not used for any experiments),a laser ?ducial?nding device,and a truth device(a device used in the simulator to extract information about the robot’s position status to record experimental data).The dimensions of the simulated robots are33x44cm.

The experiments were carried out by executing each of the four movement algorithms in?ve di?erent simulated environments with di?erent numbers of robots(10,20,30,40,and50)and two di?erent durations(5min and10min). The environments are described in Table1and shown in Figure2.In each experiment the robots started out clustered together near the center of the map,each robot facing a random direction(except for the house world where the robots began in the left most room).

Experiments were carried out within the Player/Stage[10,3]architecture. The Player Java client library was used to control the robots for all experi-ments.Each robot’s control ran on its own thread,and all robot control code was executed on the same machine as the Stage simulation.This was done

6Ryan Morlok and Maria Gini

Fig.2.The simulated worlds used in the experiments.Top:Square,Convex,Con-cave,and House worlds.Bottom:Hospital World.

World Size Key Features

Square30x30m Simplest world,large open area designed to provide a baseline

for other worlds

Convex30x30m Simple world with basic,convex obstacles;no locations where

robots can get trapped

Concave30x30m More extended,concave barriers in which robots can get

trapped if they are unwilling to backtrack

House41x16m World modeled after a simple house blueprint;robots interact

with a simpli?ed map of a real world environment

Hospital140x54m Complex world(packaged with Stage)designed to test robots

in complicated world with many corridors and rooms Table1.Simulated test environments;see Figure2for visual reference.

primarily because of the di?culty of starting all robots simultaneously on multiple machines.Each of the experiments was run four times.

Because the Stage simulator does not provide built in utilities for analyzing the performance of the robots as they observe the environment,snapshots

of relevant data were taken from the simulation for later o?-line analysis. This was done by attaching a Truth device(a simulated device in Stage)to each robot.The Truth device is a device that can either get or set the world coordinates/orientation of any object to which it is attached.For each robot, an additional thread running on the same JVM queried the robot via the Truth device and recorded its position/orientation once per second.Data for

all robots for a given experiment was written out to the same?le.The data could then be used later(in combination with the world map)to determine the observation coverage of the environment.

Dispersing robots in an unknown environment 7

For each experiment,a single,large ?le containing the position coordinates of all the robots was created.In order to determine observation coverage of the robots,the data was analyzed in combination with the world map used in the experiment.The procedure was implemented in Matlab.

First,all data points are loaded from the ?le.A binary bitmap of the same size as the world bitmap is created.Each pixel in the binary bitmap represents a discrete location in the world,which can be in one of two states:observed

(1)or not observed (0).Initially,all pixels are not observed.For each location that is taken from the robot position ?le,all of the pixels within a set radius are set to observed.This is repeated for all the locations in the position ?le.

After all locations in the position ?le are processed,the observed region is oversized.Some areas are marked as observed when they are in fact un-observed.This is because the observed region includes pixels that are in fact obstacles,as well as those that are outside the accessible region of the world (a closed region from which the robot cannot escape).To account for this,a logical AND is performed on the observed binary bitmap and the interior region of the world.The interior region of the world is found by performing a ?ood ?ll,beginning at the start location of one of the robots.This leaves the observed bitmap as the locations that have fallen within the robots’observed radii at some point,that are valid,and accessible points in the

world.

Fig. 3.A problem with the model used to calculate the re-gion observed by a robot.Note that this procedure is not a per-

fect model of what the robots could actually

observe,especially in blueprint-like environ-

ments.Consider the situation shown in Fig-

ure 3.Suppose the elongated rectangle rep-

resents a wall separating two rooms (both of

which are accessible to the robot).Here the

algorithm will mark the area beyond the wall

as observed,where clearly it is not.We devel-

oped a method to deal with this problem,but we did not use it in the data reported,since it signi?cantly increases the time for data anal-ysis and only leads to a minor improvement

in accuracy for a relatively small view radius

of the robots.Because of this,it should be noted that data reported for the House and Hospital worlds are slight overestimates of the actual values.5Experimental Results

Table 2summarizes the results from both the ?ve and ten minute experiment runs for the four algorithms in all ?ve environments with di?erent numbers of robots.The percents values in the table indicate the percentage of the accessible area of the environment that was observed by the robots.

8Ryan Morlok and Maria Gini

5Minutes10Minutes

10203040501020304050

Square Environment

Fiducial61.0%77.4%81.6%85.5%83.4%81.2%95.8%96.9%96.9%97.1% Follow wall28.8%57.4%52.8%53.8%57.3%56.6%76.1%77.0%68.4%78.6% Random walk50.2%65.7%76.4%84.8%79.8%73.3%90.6%96.6%95.8%97.7% Seek open18.7%36.7%43.9%47.3%44.4%41.8%72.7%70.7%87.4%77.0%

Convex Environment

Fiducial52.2%66.1%71.6%75.1%75.6%74.7%82.9%93.7%94.4%92.0% Follow wall22.7%41.8%36.2%36.0%37.5%28.8%64.6%61.1%59.8%55.4% Random walk44.2%58.5%62.4%64.9%65.9%69.8%82.4%84.0%88.2%93.1% Seek open18.9%30.2%40.9%38.4%33.6%36.6%52.8%59.9%65.9%56.9%

Concave Environment

Fiducial46.2%58.7%64.5%67.3%69.0%67.8%85.9%85.8%90.7%88.5% Follow wall14.9%34.6%37.1%35.7%35.5%35.5%53.1%58.3%52.4%56.3% Random walk33.8%48.2%56.2%64.8%59.8%51.6%73.4%78.6%79.0%86.4% Seek open16.2%29.4%35.0%33.5%40.8%36.1%49.1%53.4%54.6%60.9%

Home Environment

Fiducial37.0%40.0%43.0%40.9%40.7%39.7%46.3%47.2%44.5%43.9% Follow wall23.9%22.3%27.3%30.9%35.2%31.4%32.6%39.1%37.0%40.7% Random walk33.3%37.1%40.0%38.6%40.4%39.2%41.6%42.9%44.4%44.1% Seek open23.8%31.6%33.6%33.4%35.2%35.5%37.6%36.6%37.1%36.9%

Hospital Environment

Fiducial 5.6% 6.0%7.0%7.7%8.7%8.4%11.0%10.6%10.3%13.3% Follow wall 3.3% 3.6% 3.1% 3.7% 4.3% 4.1% 6.5% 5.5%7.0% 5.0% Random walk 4.7% 4.4% 4.1% 6.1% 5.9% 5.0% 6.5% 4.9%7.5%8.0% Seek open 3.4% 3.5% 3.7% 3.9% 4.8% 3.4% 3.5% 3.8% 4.3% 5.1%

Table2.Results for all experiments

The data shows that the Fiducial algorithm performs the best in every situation.This is not surprising in that this algorithm has access to more data than the other algorithms.It does,however,indicate that knowledge of the locations of the other robots can help to speed up the exploration process.

The results for the Fiducial algorithm can be seen graphically in Fig-ure4.This illustration shows that the robots had di?culty observing the more enclosed areas of the map,which is to be expected since the obstacle avoidance portion of the movement algorithm tends to favor regions in which it does not run into things.To get into the enclosed-hook region,the robot would have to intersect the obstacle,and then by random chance be redirected towards the hook region.None of the movement algorithms have the ability to be naturally attracted toward this type of region.

Figure4also shows the performance of the Fiducial algorithm in the House world.All robots started out in the garage(the large,left most rectangle that is all green(light colored))clustered together,facing di?erent,random directions.Here we can see that the robots managed to make it into the house,

Dispersing robots in an unknown environment9

Fig. 4.Results of the Fiducial algorithm in the concave,house,and hospital worlds.Red(dark)indicates an unexplored area,green(light)indicates an observed region.The concave and house results were obtained by running50robots for10 minutes,and the hospital results were obtained by running50robots for1hour. but not past the?rst room.Doors are a natural obstacle for all movement algorithms in this experiment,in that they present a situation where there is obstacle on both sides,with a small opening.This can cause the avoidance logic to move away from the door area when it is encountered,unless the robot hits the door dead on.

What was somewhat surprising was that the RandomWalk movement algorithm performed second best among those algorithms tested,and generally close to the performance of the Fiducial movement algorithm.The similarity in performance between them should be expected,since for the majority of the time,both are acting in the same manner.The?ducial algorithm will only act di?erently than the random walk algorithm early in the simulation,when all of the robots are still clustered together.After the robots have spread out,the movement patterns become identical.The Fiducial algorithm simply speeds this spreading process.

Both FollowWall and SeekOpen have some innate?aws.The?aw in FollowWall is that it assumes two things.First,it assumes the robot to be in a closed region with no internal,isolated obstacles.Robots using FollowWall are likely to?nd an obstacle and orbit it inde?nitely.The robot would require some form of self-position estimation to detect when it traverses the same positions over and over again.The second problem with FollowWall has been mentioned previously.The algorithm has no ability to distinguish between robots and obstacles(namely walls)because of this,in the initial robot cluster,many robots ended up trying to follow each other as walls,and ended up going in circles.

The fundamental problem with the SeekOpen algorithm is that it is sub-ject to orbiting around local maxima for“openness.”Some robots using this algorithm were observed to be traveling in small circles,remaining in one area of the map.SeekOpen also tends to prevent robots from going through nar-row passageways such as doors.The hook in the Concave world,as shown, could act as one such narrow passageway.

10Ryan Morlok and Maria Gini

6Conclusions and Future Work

We have examined the performance of several movement algorithms at dis-persing a group of robots in an unknown environment when starting from a initial cluster.The algorithms have been tested in various virtual worlds varying from a simple open area to a complex real-world building.The results show that even approximate knowledge of the locations of close-by robots helps the robots to spread out.The next step is to move into the real world. The combination of these algorithms and a group of small robots would make for an e?ective system for automatically exploring unknown worlds. Acknowledgments

Work supported in part by NSF under grants EIA-0224363and EIA-0324864.

References

1.E.Bonabeau,M.Dorigo,and G.Theraulaz.Swarm Intelligence:From Natural

to Arti?cial Systems.Oxford University Press,1999.

2.H.Choset.Coverage for robotics-a survey of recent results.Annals of Mathe-

matics and Arti?cial Intelligence,31:113–126,2001.

3.B.P.Gerkey,R.T.Vaughan,K.St¨o y,A.Howard,G.S.Sukhatme,and M.J.

Matari′c.Most valuable player:A robot device server for distributed control.

In Proc.IEEE/RSJ Int’l Conf.on Intelligent Robots and Systems,pages1226–1231,Oct.2001.

4.T.-R.Hsiang,E.Arkin,M.A.Bender,S.Fekete,and J.Mitchell.Algorithms

for rapidly dispersing robot swarms in unknown environments.In Proc.5th Workshop on Algorithmic Foundations of Robotics(WAFR),2002.

5.T.-R.Hsiang,E.Arkin,M.A.Bender,S.Fekete,and J.Mitchell.Online dis-

persion algorithms for swarms of robots.In Proc.of the19th Annual ACM Symposium on Computational Geometry(SoCG),pages382–383,2003.

6.S.Koenig,B.Szymanski,and Y.Liu.E?cient and ine?cient ant coverage

methods.Annals of Mathematics and Arti?cial Intelligence,31:41–76,2001. https://www.wendangku.net/doc/e516576440.html,tombe.Robot Motion Planning.Kluwer Academic Publ.,Norwell,MA,

1991.

8.D.Payton,M.Daily,R.Estkowski,M.Howard,and C.Lee.Pheromone robotics.

Autonomous Robots,11(3):319–324,Nov2001.

9.P.E.Rybski,S.A.Stoeter,M.Gini,D.F.Hougen,and N.Papanikolopou-

los.Performance of a distributed robotic system using shared communications channels.IEEE Trans.on Robotics and Automation,22(5):713–727,Oct.2002.

10.R.T.Vaughan.Stage:A multiple robot simulator.Technical Report IRIS-

00-394,Institute for Robotics and Intelligent Systems,University of Southern California,2000.

11.I.A.Wagner,M.Lindenbaum,and A.M.Bruckstein.MAC vs PC–deter-

minism and randomness as complementary approaches to robotic exploration of continuous unknown domains.Int’l Journal of Robotics Research,19(1):12–31, 2000.

雅思历年真题口语题目汇总

雅思历年真题口语题目汇总 version 01old person describe an old man influenced you 1.who was he 2.when did you know him 3.what he did and explain why he influenced you part3 1.老人的经验有什么问题存在? 2.喜欢什么艺术品? 3.给老人拍照片时候注意什么呢? 4.你们国家对老年人是什么态度？ 5.你认为这个社会在哪些方面对老年人不太好？ 6.老人在你们家有什么影响？ 7.你认为老年人在看问题的时候跟年轻人有什么不一样？ 8.他们对大家有什么影响？ version 02 city 1.where it is located? 2. what special for you? 3. why you want to stay there? part 3 1.please compare 100 hundred years old city and modern city and what predict about the city in the future. 2.上海是个怎样的城市 3.都有那些著名建筑

4.你想为这个城市做些什么？ 5.有哪些现象有待提高或者那些提倡 version 03 room part2: 1.what's your favorite room in your home 2.what it likes you live 3.what you do in the room normally and explain why you like it part3: 1.你认识你的邻居吗？ 2.城市里的房子和乡村有什么不同？ 2003年9月换题后的口语topic Old person Describe a older person you know You should say:Who he or she is How you know him or her How he or she is And explain what infection he or she give you and in what aspect Further question： 1、你们国家对老年人是什么态度？ 2、你认为这个社会在哪些方面对老年人不太好？ 3、老人在你们家有什么影响？ 4、你认为老年人在看问题的时候跟年轻人有什么不一样？ 5、他们对大家有什么影响？

完整版初中英语语法大全知识点总结

英语语法大全 初中英语语法 学习提纲 一、词类、句子成分和构词法： 1、词类：英语词类分十种： 名词、形容词、代词、数词、冠词、动词、副词、介词、连词、感叹词。 1、名词(n.)：表示人、事物、地点或抽象概念的名称。如：boy, morning, bag, ball, class, orange. ：who, she, you, it . 主要用来代替名词。如)： 2、代词(pron.3、形容词(adj..)：表示人或事物的性质或特征。如：good, right, white, orange . 4、数词(num.)：表示数目或事物的顺序。如：one, two, three, first, second, third, fourth. 5、动词(v.)：表示动作或状态。如：am, is,are,have,see . 6、副词(adv.)：修饰动词、形容词或其他副词,说明时间、地点、程度等。如：now, very, here, often, quietly, slowly. 7、冠词(art..)：用在名词前,帮助说明名词。如：a, an, the. 8、介词(prep.)：表示它后面的名词或代词与其他句子成分的关系。如in, on, from, above, behind. 9、连词(conj.)：用来连接词、短语或句子。如and, but, before . 10、感叹词(interj..)表示喜、怒、哀、乐等感情。如：oh, well, hi, hello. 2、句子成分：英语句子成分分为七种：主语、谓语、宾语、定语、状语、表语、宾语补足语。 1、主语是句子所要说的人或事物,回答是“谁”或者“什么”。通常用名词或代词担任。如：I'm Miss Green.(我是格林小姐) 2、谓语动词说明主语的动作或状态,回答“做(什么)”。主要由动词担任。如：Jack cleans the room every day. (杰克每天打扫房间) 3、表语在系动词之后,说明主语的身份或特征,回答是“什么”或者“怎么样”。通常由名词、 代词或形容词担任。如：My name is Ping ping .(我的名字叫萍萍) 4、宾语表示及物动词的对象或结果,回答做的是“什么”。通常由名词或代词担任。如：He can spell the word.(他能拼这个词) 有些及物动词带有两个宾语,一个指物,一个指人。指物的叫直接宾语,指人的叫间接宾语。间接 宾语一般放在直接宾语的前面。如：He wrote me a letter . (他给我写了 一封信) 有时可把介词to或for加在间接宾语前构成短语,放在直接宾语后面,来强调间接宾语。如：He wrote a letter to me . (他给我写了一封信)

初中英语中主谓一致详解

主谓一致详解 【基础知识】 主谓一致指“人称”和“数”方面的一致关系。对大多数人来说，往往会在掌握主语和随后的谓语动词之间的一致问题上遇到困难。一般情况下，主谓之间的一致关系由以下三个原则支配： 语法一致原则(grammatical concord) 意义一致原则(notional concord) 就近原则(principle of proximity) （一）语法一致原则 用作主语的名词词组中心词和谓语动词在单、复数形式上的一致，就是语法一致。也就是说，如果名词中心词是单数，动词用单数形式；如果名词中心词是复数，动词用复数形式。例如： This table is a genuine antique. Both parties have their own advantages. Her job has something to do with computers. She wants to go home. They are divorcing each other. Mary was watching herself in the mirror. The bird built a nest. Susan comes home every week-end. （二）意义一致原则 有时，主语和谓语动词的一致关系取决于主语的单、复数意义，而不是语法上的单、复数形式，这样的一致关系就是意义一致。例如： Democratic government gradually take the place of an all-powerful monarchy. A barracks was attacked by the guerilla. Mumps is a kind of infectious disease. The United States is a developed country. It is the remains of a ruined palace. The archives was lost.

雅思口语真题

IELTS SPEAKING TEST Part 1 Work or study 1.Do you work or are you a student? 2.What subject(s) are you studying? 3.Why did you choose to study that subject? 4.(For high school) Why did you choose those subjects? 5.Is that a popular subject (to study) in your country? 6.Do you think it's popular because people want to gain knowledge or is there some other reason? 7.What are the most popular subjects (= university degree courses = majors) in China? 8.Did your family help you choose that course? 9.What school (or university) do you go to? 10.Why did you choose that university (or, school)? 11.Do you have any recreational or entertainment activities at your school/university? 12.How do you like your subject? 13.What do you like about your subject? 14.Do you think it's important to choose a subject you like? (Why?) 15.(For university students) What are your favourite classes/ courses/ subjects at university? 16.(For high school students) What's your favourite subject at school? 17.Is your subject interesting? (Why? / Why not?) 18.Do you think it's important to choose a subject you are interested in? (Why?) 19.What's the most interesting part of your subject? 20.(For high school) What's the most interesting of your subjects at school? 21.What are your future work plans? (after you graduate) 22.Do you think what you are studying now will be very useful (or, relevant or, important) for this type of work? 23.How will you (or, how do you plan to) get the job you want? 24.Do you think it will be easy to find that kind of work? 25.Why are you taking the IELTS test? 26.In addition to gaining knowledge, what other ways have you benefited from your school/university experience? Accommodation 1.What kind of housing/accommodation do you live in? 2.Who do you live with? 3.How long have you lived there? 4.Do you plan to live there for a long time? 5.Can you describe the place where you live? 6.Which room does your family spend most of the time in? 7.What do you usually do in your house/flat/room? 8.Are the transport facilities to your home very good? 9.Do you prefer living in a house or a flat? 10.Please describe the room you live in. 11.Is there anything you don't like about the place where you live?

雅思历年真题口语题目汇总

雅思历年真题口语题目汇总 version01old person describe an old man influenced you 1.who was he 2.when did you know him 3.what he did and explain why he influeced you part3 1.老人的经验有什么问题存在? 2.喜欢什么艺术品? 3.给老人拍照片时候注意什么呢? 4.你们国家对老年人是什么态度？ 5.你认为这个社会在哪些方面对老年人不太好？ 6.老人在你们家有什么影响？ 7.你认为老年人在看问题的时候跟年轻人有什么不一样？ 8.他们对大家有什么影响？ version02city 1.where it is located? 2.what special for you? 3.why you want to stay there? part3 1.please compare100hundred years old city and modern city and what predict about the city in the futu re. 2.上海是个怎样的城市 新东方批改网（https://www.wendangku.net/doc/e516576440.html,）,在线雅思作文批改，雅思口语批改。语法纠错、恶补，制定考试计划，

3.都有那些著名建筑 4.你想为这个城市做些什么？ 5.有哪些现象有待提高或者那些提倡 version03room part2: 1.what's your favorite room in your home 2.what it likes you live 3.what you do in the room normally and explain why you like it part3: 1.你认识你的邻居吗？ 2.城市里的房子和乡村有什么不同？ 2003年9月换题后的口语topic Old person Describe a older person you know You should say:Who he or she is How you know him or her How he or she is And explain what infection he or she give you and in what aspect Further question： 1、你们国家对老年人是什么态度？ 2、你认为这个社会在哪些方面对老年人不太好？ 3、老人在你们家有什么影响？ 新东方批改网（https://www.wendangku.net/doc/e516576440.html,）,在线雅思作文批改，雅思口语批改。语法纠错、恶补，制定考试计划，

六年级英语语法知识点汇总

六年级语法总复习 一、词汇 （一）一般过去时态 一般过去时态表示在过去的某个时间发生的动作或存在的状态，常和表示过去的时间状语连用。例如yesterday, last weekend ,last Saturday ,等连用。基本句型：主语+动词的过去式+其他。例句——What did you do last weekend?你上周做什么了？ ——I played football last weekend.我踢足球了。 ★规则动词过去式的构成 ⒈一般在动词原形末尾加-ed。例如：play—played ⒉词尾是e的动词直接加-d。例如：dance—danced ⒊末尾只有一个辅音字母的重读闭音节词，先双写这个辅音字母，再加-ed。例如stop（停止）--stopped ⒋结尾是“辅音字母+y”的动词，变“y”为“i”，再加-ed，例如：study--studied ★一些不规则变化的动词过去式 am/is—was are—were go—went swim—swam fly—flew do—did have—had say—said see—saw take—took come—came become—became get—got draw—drew hurt—hurt read—read tell—told will—would eat—ate take—took make—made drink—drank sleep(睡觉)—slept cut（切）--cut sit(坐)—sat begin(开始)—began think—thought find—found run(跑)---ran buy—bought win—won give(给)—gave sing—sang leave—left hear（听）--heart wear—wore （二）一般现在时态 一般现在时态表示包括现在时间在内的一段时间内经常发生的动作或存在的状态，表示习惯性或客观存在的事实和真理。常与often ,always ,usually ,sometimes ,every day等连用。基本句型分为两种情况： ●主语（非第三人称）+动词原形+其他。例句：——What do you usually do on the weekend?——I usually do my homework on the weekend. ●主语（第三人称）+动词的第三人称单数形式+其他。例句： ——What does Sarah usually do on the weekend?萨拉通常在周末干什么？ ——She usually does her homework on the weekend.她通常在周末做她的家庭作业。 ★动词第三人称单数形式的变化规则 ⒈一般直接在动词词尾加-s.例如：play—plays ⒉以s ,x ,ch,sh结尾的动词加-es。例如：watch—watches ⒊以辅音字母加y结尾的动词，变y为i，再加es,例如：fly—flies ⒋个别不规则变化动词，需单独记忆，例如：do—does go—goes （三）现在进行时态 现在进行时态表示说话人现在正在进行的动作。基本句型：主语+be+动词的-ing+其他。 例如:——What are you doing ?你在干什么？ ——I am doing my homework..我正在做作业。 ★动词现在分词的变化规则 ⒈一般直接在词尾加ing ，例如;wash—washing ⒉以不发音e字母结尾的动词，去掉e ,再加ing.例如：make—making ⒊末尾只有一个辅音字母的重读闭音节词，要双写最后一个辅音字母再加ing.例如swim—swimming （四）一般将来时态 一般将来时态表示将来某一时间或某一段时间内发生的动作或存在的状态。常与表示将来的时间如tomorrow ,next weeken ,this afternoon 等连用。我们通常用will,be going to+动词原形来表示一般将来时态。

l主谓一致讲解最全面主谓一致讲解

主谓一致的讲解 主谓一致是指： 1）语法形式上要一致，即名词单复数形式与谓语要一致。 2）意义上要一致，即主语意义上的单复数要与谓语的单复数形式一致。 一、并列结构作主语时的主谓一致 1.由and 连接主语时 And 连接的两个或多个单数可数名词、不可数名词或代词作主语时根据意义或概念确定谓语用单数或复数 1)并列主语表示不同的人、物或概念时谓语动词用复数 Li Ming and Zhang Hua are good students. Like many others, the little tramp and the naughty boy have rushed there in search of gold. 小流浪汉和调皮的小男孩也赶到那里寻找金子 Both rice and wheat are grown in this area. 2)并列主语表示同一个人、物或概念时，谓语动词用单数形式。 The professor and writer is speaking at the meeting. 那位教授兼作家正在会上发言 A journalist and authour lives on the sixth floor. 一位新闻记者兼作家 His lawyer and former college friend was with him on his trip to Europe. 他的律师兼大学时代的朋友陪他去欧洲旅行 The Premier and Foreign Minister was present at the state banquet. 总理兼外长 比较：the writer and the educator have visited our school. the writer and educator has visited our school. His lawyer and his former college friend were with him on his trip to Europe. 注意：指同一个人或物时，并列主语前只用一个冠词，指不同的需要分别加冠词，但两个名词具有分别的对立的意思时只需要一个冠词即可 A boy and girl are playing tennis. 3)并列主语前有each, every, many a , no 等修饰时谓语动词用单数 Each doctor and (each) nurse working in the hospital was asked to help patients. Every man, woman and child is entitled to take part in the activity. 有权参加 Every boy and (every) girl admires him for his fine sense of humour. Many a boy and (many a ) girl has made the same mistake No boy and no girl is there now.没有任何男孩和女孩在那里 注意：many a 跟单数可数名词但是表示复数意义翻译为很多 Many a student was disappointed after seeing the movie. 4)并列主语为不可分的整体时，谓语动词用单数 A law and rule about protecting environment has been drawn up. 关于保护环境的法律法规已经起草完成。 The knife and fork has been washed 刀叉已经被洗好 War and peace is a constant theme in history 战争与和平是历史永恒的主题 注意;常被视为主体的结构 A cup and saucer 一副杯碟 A horse and cart 马车 A knife and fork 一副刀叉

雅思口语部分PART1附答案(1)解析

PART1 Names 1.Who gave you your name? The name was given by my parents. 2.Does your name have any particular (or special meaning? I don’t know about that. 3.Do you like your name? I guess yes. 4.In your country do people feel that their name is very important? Most of people think so, because names are the first impression to other people. 5.Would you like to change your name? I don’t think it’s necessary. 6.Is it easy to change your name in your country? Maybe there will be some procedures to do. 7.Who usually names babies in your country? Parents or grandparents. 8.Do you have any special traditions about naming children? It depends on the tradition of the hometown. 9.What names are most common in your hometown?

【雅思口语预测】雅思口语话题题库范文

【雅思口语预测】雅思口语话题题库范文 雅思口语精准预测刘薇简介环球雅思北京学校副校长最受欢迎口语教师授课时激情中透着温婉幽默里渗入励志秉承让学生 "轻松复习事半功倍"的教学理念。 从事雅思教学多年北京大学传播学硕士曾学术访问哈佛大学 耶鲁大学等世界顶尖名校。 著作《雅思九天口语高分之路》、《剑桥雅思全真试题原版解析8》。本月预测重点本预测适用于xx年4月雅思考试超高频和重回题库考题、一级重点、都是重点类别需要考生准备细致答案。二级重点考生考生准备思路即可。 Part 1 分类汇总一基本信息类 1. Name 什么名字，中文名含义，未来换吗 2. Study or work 上班还上学上学问专业或最喜欢学科上班工作内容，参加过培训吗，未来换工作吗 3. Hometown 家乡哪里，有什么特别的地方，未来想居住的地方 4. Weather 喜欢什么天气家乡天气，家乡天气今年的变化 5. Family 家庭情况，家庭时间如何度过，全家人喜欢一起做什么二衣食住行类 1. Living 住flat还是house，最喜欢的房间，如果重装修会怎么做 2. Building 喜欢什么建筑，你们国家的人更喜欢传统还是时尚建筑，未来建筑的发展 3.

Countryside 喜欢郊区吗，郊区与城市的区别，什么人喜欢郊区 4. Transport 家乡交通特点，公共交通和私人交通的优缺点，未 来如何改善 5. Boating 划过船吗，坐过船吗，好处，你们国 家水运多吗 6. Clothes and fashion 喜欢什么衣服，喜欢时 尚吗，时尚对生活的影响 7. Bags 平时背包包吗，什么类型的 宝宝，知道哪些包包，丢过包包吗 8. Keeping healthy and fit 如何保持健康，如何保持身材 9. Daily routine and favorite time in a day 典型的一天，一天中最喜欢的时间段 10. Sleeping 几点睡几点起，睡不着怎么办，如何培养好的睡眠习惯 三娱乐休闲 1. Music 喜欢什么类型音乐，喜欢在家听还是concert hall。会乐器吗，什么乐器。 2. dancing 会舞蹈吗，中国人喜欢什么舞蹈，传统舞蹈得到了 传承吗 3. photography 喜欢摄影吗，喜欢给别人拍还是别人 给你拍，喜欢收藏照片吗 4. painting 会画画儿吗，小孩子学 习绘画的好处，有什么绘画技巧吗 5. sport 最喜欢什么运 动，好处 6. museum 喜欢去博物馆吗，好处，博物馆该收费吗7. leisure and relaxation 喜欢什么休闲活动，跟谁，在哪 儿，好处，中国人都喜欢什么休闲 8. weekends 如何度过周 末，周末加班吗，老板应该给发工资吗 9. postcards 喜欢 ___吗，发给别人或者收到过吗，跟email的区别 10. public park 喜欢去公园吗，做什么，应该有更多公园吗四科技生活

2019雅思口语考试真题(9)

2019年9-12月雅思口语Part 2新题预测：有意义的 歌 有意义的歌 P2 Describe a song that is meaningful to you. You should say： what song it is; when you first heard it; what the song is about; and explain why it is meaningful to you. P3 1.Is it good for children to learn music? 2.How do Chinese people learn music? 3.Do men and women play different kinds of musical instruments? 4.what's the status of traditional music now? 5.Do all grades have music class? 6.Is music class important? 7.What kind of music do children like? 8.Why would some parents force their children to learn some musical

instruments? 解析 题目要求考生描述“一首对你有意义的歌”。作答要点包括：是什么歌、什么时候第一次听到、歌是关于什么的，并说明为什么这首歌对你有特殊意义。 范文 Three years ago, a song came out that celebrated being young and relaxed and I was instantly crazy about it. It was called ‘Young for you’ and the beat and melody were off the hook! The song became a massive hit across China that year and you could hear it being played everywhere. When I listen to the song now it immediately takes me back to my youth. I first heard the song playing in a shopping mall and was instantly drawn to it. The singer's voice was like no other I had ever heard before and the melody was haunting and sad. I stopped shopping and used an app on my phone to identify the tune. I immediately downloaded the album off iTunes.

小学英语语法知识点汇总!

小学英语语法知识点汇总！ 01 人称代词 主格：I we you she he it they 宾格：me us you her him it them 形容词性物主代词：my our your her his its their 名词性物主代词：mine ours yours hers his its theirs 02 形容词和副词的比较 (1) 一般在形容词或副词后+er older ,taller, longer, stronger (2) 多音节词前+more more interesting, etc. (3) 双写最后一个字母，再+er bigger fatter, etc. (4) 把y变i，再+er heavier, earlier (5) 不规则变化： well-better, much/many-more, etc. 03 可数词的复数形式 Most nouns + s abook –books

Nouns ending in aconsonant +y - y+ ies a story—stories Nouns ending in s,sh, ch or x + es a glass—glasses a watch-watches Nouns ending in o+s or +es a piano—pianos a mango—mangoes Nouns ending in for fe - f or fe +ves a knife –knives a shelf-shelves 04 不可数名词（单复数不变） bread, rice, water ,juice等。 05 缩略形式 I’m= I a，you’re = you are，she’s= she is，he’s = he is it’s= it is，who’s =who is，can’t =can not，isn’t=is not等。 06 a/an a book, a peach an egg，an hour 07 Preposition on, in ,in front of, between, next to, near, beside, at,behind. 表示时间：at six o’clock, at Christmas, at breakfast

高中英语主谓一致知识点讲解

高中英语主谓一致知识点讲解 本文主要讲解主谓一致，并列结构作主语时谓语用复数主谓一致中的靠近原则谓语动词与前面的主语一致 等常见考点。 主谓一致是指： 1）语法形式上要一致，即单复数形式与谓语要一致。 2）意义上要一致，即主语意义上的单复数要与谓语的单复数形式一致。 3）就近原则，即谓语动词的单复形式取决于最靠近它的词语， 一般来说，不可数名词用动词单数，可数名词复数用动词复数。例如： There is much water in the thermos. 但当不可数名词前有表示数量的复数名词时，谓语动词用复数形式。例如： Ten thousand tons of coal were produced last year. 并列结构作主语时谓语用复数，例如： Reading and writing are very important. 读写很重要。 注意：当主语由and连结时，如果它表示一个单一的概念，即指同一人或同一物时，谓语动词用单数，and 此时连接的两个词前只有一个冠词。例如： The iron and steel industry is very important to our life. 钢铁工业对我们的生活有重要意义。 典型例题 The League secretary and monitor ___ asked to make a speech at the meeting. A. is B. was C. are D. were 答案B. 注：先从时态上考虑。这是过去发生的事情应用过去时，先排除A.，C。本题易误选D，因为The League secretary and monitor 好象是两个人，但仔细辨别，monitor 前没有the，在英语中，当一人兼数职时只在第一个职务前加定冠词。后面的职务用and 相连。这样本题主语为一个人，所以应选B。

雅思口语考试形式及内容全面解析

雅思口语考试形式及内容全面解析 雅思口语考试是考生与考官之间进行一对一交流的形式，考官对考生的英语口语水平进行考察。雅思口语考试分为三个部分，考生可以以此使用不同的口语表达技能。雅思考试口语部分将被录音。下面就和大家分享，来欣赏一下吧。 雅思口语考试形式及内容全面解析 雅思口语Part 1 雅思口语考试形式：考官会向考生进行自我简介，并核对考生的身份。之后，考官会就考生熟悉的话题(如朋友、兴趣习惯或者食物) 进行询问。为保证题目的一致性，这些问题都是从一个事先拟定的范围内抽取的。 考试时间有多长：4-5分钟。 这部分考察的是什么技能：这部分考察的是考生就日常性的观点和信息、常见的生活经历或情形以回答问题的形式进行交流的能力。 雅思口语Part 2 雅思口语考试形式：这部分为考生作个人陈述。考官会交给考生一个答题任务卡、铅笔和草稿纸做笔记。答题任务卡上会给

出一个话题和需要在个人陈述中包含的要点，并在最后提示考生解释这个话题的某一个方面。有效地使用答题任务卡上的提示可以帮助考生思考讲述的话题、组织内容、并持续地陈述2分钟时间。在准备时间内做一些笔记也可以帮助考生安排好陈述的结构。考生有一分钟的准备时间，之后考官会要求考生就相关内容讲述1-2分钟。考官会在2分钟后打断考生，并在最后提问一两个问题作为结束语。 考试时间有多长：3-4分钟。 这部分考察的是什么技能：这部分考察得是考生(在没有任何其它提示的情况下)就一个特定的话题进行较长时间的陈述的能力，考察考生是否能恰当地运用语言、是否能连贯地组织自己的观点。考生有可能需要联系自己的经历来完成这部分内容。 雅思口语Part 3 雅思口语考试形式：在这部分考试中，考官和考生将对第二部分中涉及的话题进行讨论，讨论将为更加广泛和抽象，在恰当的时候还会更加深入。 考试时间有多长：4-5分钟。 这部分考察的是什么技能：这部分考察的是考生表达和论述看法、分析、讨论以及深入思考问题的能力。 雅思口语范文：describe something made you laugh

剑桥雅思9口语真题+解析Test3-Part3

剑桥雅思9口语真题+解析Test3-Part3 摘要：剑桥雅思9口语真题+解析Test3-Part3，剑桥雅思9Test3口语Part3讨论话题举例：Reasons for daily travel，有需要的同学抓紧时间下载吧! Discussion topics: Reasons for daily travel Example questions: · Why do people need to travel every day? 为什么人们每天都需要出行? 笨鸟雅思口语名师点题 本题询问人们为什么每天穿梭于各地。通常针对询问people的问题，可以相应列举出不同人群，如上班的人、上学的人、走亲访友的人、旅游的人等，同时针对某类人群适当进行拓展。 高分示例 People need to travel for lots of different reasons. Almost everybody has to travel to work, so they can earn money and provide for their families. If a person is lucky, then they will live very close to the place where they work, so they can travel on foot or by bicycle. Sometimes, however, people live very far away, so they must travel by bus, train, car, or even sometimes by plane or boat! Lots of people also travel to see friends and family members, who might live in a different city, or even a different country. But often people just like to travel for fun, to go to somewhere new, or to see a famous tourist site or place of natural beauty. 参考译文 人们有各种各样的理由来来往往。几乎每个人都要去上班，这样才能赚钱养家。幸运的人住的地方离工作单位很近，可以步行或者骑车去上班。但是，有时候，住的地方比较远的人就必须坐公交车、坐火车、开车，甚至坐飞机、坐船去上班。大部分人都会去异地甚至异国拜访朋友、看望家人。但是人们经常是出去玩，去探索新的地方，去看看著名的景点，去看看自然美景。 亮点表达 earn money 赚钱 tourist site 旅游景点 provide for 供养

初三英语语法知识点归纳

初中英语语法速记口诀大全 很多同学认为语法枯燥难学，其实只要用心并采用适当的学习方法，我们就可以愉快地学会英语，掌握语法规则。笔者根据有关书目和多年教学经验，搜集、组编了以下语法口诀，希望对即将参加中考的同学们有所帮助。 一、冠词基本用法 【速记口诀】 名词是秃子，常要戴帽子， 可数名词单，须用a或an， 辅音前用a，an在元音前， 若为特指时，则须用定冠， 复数不可数，泛指the不见， 碰到代词时，冠词均不现。 【妙语诠释】冠词是中考必考的语法知识之一，也是中考考查的主要对象。以上口诀包括的意思有：①名词在一般情况下不单用，常常要和冠词连用；②表示不确指的可数名词单数前要用不定冠词a或an，确指时要用定冠词the；③如复数名词表示泛指，名词前有this，these，my，some等时就不用冠词。 二、名词单数变复数规则 【速记口诀】 单数变复数，规则要记住， 一般加s，特殊有几处： 末尾字母o，大多加s， 两人有两菜，es不离口， 词尾f、fe，s前有v和e； 没有规则词，必须单独记。 【妙语诠释】①大部分单数可数名词变为复数要加s，但如果单词以/t?蘩/、/?蘩/、/s/发音结尾(也就是单词如果以ch，sh，s，x等结尾)，则一般加es；②以o结尾的单词除了两人(negro，hero)两菜(tomato，potato) 加es外，其余一般加s；③以f或fe结尾的单词一般是把f，fe变为ve再加s；④英语中还有些单词没有规则，需要特殊记忆，如child—children，mouse—mice，deer—deer，sheep—sheep，Chinese—Chinese，ox—oxen，man—men，woman—women，foot—feet，tooth —teeth。 三、名词所有格用法 【速记口诀】 名词所有格，表物是“谁的”， 若为生命词，加“’s”即可行， 词尾有s，仅把逗号择； 并列名词后，各自和共有， 前者分别加，后者最后加； 若为无生命词，of所有格， 前后须倒置，此是硬规则。 【妙语诠释】①有生命的名词所有格一般加s，但如果名词以s结尾，则只加“’”；②并列名词所有格表示各自所有时，分别加“’s”，如果是共有，则只在最后名词加“’s”；③