Braitenberg vehicles are a widely used model of animal behaviour in robotics and Artificial Life. This paper presents the first comprehensive and formal analysis of the behaviour of Braitenberg vehicles 2 and 3. After a review of their intuitive behaviour we present their models as dynamical systems, that under circularly symmetric stimuli can be simplified and analysed using the phase plot technique. We prove that intuitive understanding is not enough to determine the potential behaviour of vehicles 2b and 3a and, under certain circumstances, they could behave as vehicles 2a and 3b respectively.
|Published (Last):||18 May 2016|
|PDF File Size:||6.39 Mb|
|ePub File Size:||20.1 Mb|
|Price:||Free* [*Free Regsitration Required]|
Not for the amusement: science fiction in the service of science. Or just science, if your agree that fiction is a part of it, always was, and always will be as long as our brains are only minuscule fragments of the universe, much too small to hold all the facts of the world, but not too idle to speculate about them. Valentino Braitenberg. This project aims to simulate some primitive robots which are given the name "Braitenberg Vehicles" , after first being stated in Vehicles: Experiments in Synthetic Psycology.
Before giving information about the simulator and the project, it is better to have some idea about the book and its enthralling author, Valentino Braitenberg. Valentino Braitenberg was born in , Bolzano. Being a German neuro-scientist and cybernetics, he is a specialist in communication and movement control. Furthermore, he deals with the anatomy of the nervous system, which leads him to write his revolutionary book Vehicles.
Before Vehicles , he had also written a number of articles concerning artificial intelligence, cybernetics and cognitive sciences. But he is most famous for Vehicles. For the reason that he considers animal brains as "pieces of computing machinery", he had created the world of vehicles like they were real animals in their own nature. He aimed to illustrate the fundamental concept of internal structure of living mechanisms, using the connections and systems used in vehicles.
Having very simple hardware, this armada of creatures show amazing complex behaviours in their world, therefore they have been used for research tools in Computational Neuroscience. Now Prof. Synthetic psychology is a branch in science, where biological behavior is synthesized instead of being analyzed.
Braitenberg is supposed to be the father of this branch, since his researches, articles and publications ushered a new era in this field. The book became consequential not only for cybernetic psychologists, but also for future work in artificial life and the coalescence of computational AI.
The discussion in the book and experiments about vehicles are leaded by "law of uphill analysis and downhill invention" which make all the stuff in the book more comprehensible, hence it is more difficult to grasp a working mechanism or system just by looking externally than to compose it up from scratch. In the light of this method, Braitenberg creates an almost-real world with the vehicles as creatures of this toy world.
Braitenberg Vehicles are actually simple automatons, showing the abilities of reactive agents, thus setting a base for artificial life. Materially, these vehicles are nothing but some combination of sensors, motors and wiring. But the real excitement is how things as basic as those gather and make up an environment that is used to model our complex world. Like humans of reality, vehicles have the property of interacting with their environment. That is the most important reason of the analogy between "them" and "us", according to the development of vehicles.
Figure 1 First generation of vehicles start with only one sensor and one motor connected with a wire Figure 1. Then, even increasing motor and sensor amount to two gets things complicated. Having combinations of different wiring and two types of sensors, integrates vehicles emotions!
Braitenberg experiments the purest form of fear, aggression and even love in his vehicles, as will be explained later. Braitenberg's toy world contains things other than vehicles, as it simulates the actual world.
Of course it is nonsense to compare a world including only creatures with the real one. The toy world is imagined to have everything that a sensor of a vehicle can react, light is the most common example that is used in the book. Vehicles behave differently in the way their hardware designed, when they see the light or another. Also, their behaviour changes by influence of alternating luminance of light. The most primitive vehicle, that I mentioned a moment ago, has a sensor-motor relation applying that velocity of motor is directly proportional to temperature sensed.
In other words, it - let's call it Vehicle1 as in the book - accelerates when it's hot, and decelerates when it gets colder. So it stays longer in cold places and even can come to rest if it is so cold to reduce the motor force below friction force.
Braitenberg concludes the creation of this vehicle with the idea of calling this creature as alive , since a dead material is not able to like cold places. From now on I will explain different kinds of vehicles briefly, since you will understand their behaviour better when you use the simulator. To move on, Vehicle1 will be upgraded by dublicating motor and sensor number, which is like biologically wrong dublication and forming of conjoined twins.
As we have two motors and two sensors now, we have two different combination to wire them To wire both to both is exactly Vehicle1. First think about connecting ones at same sides, calling Vehicle2a Figure 2. Then what happens, if right sensor sees less light than left one, right motor will turn less than left and the vehicle itself will turn right, where the light source is not at. This means Vehicle2a will always move to the orientation that its source doesn't exist.
If the sensors sense the light in the same amount, Vehicle2a will accelerate towards it and hit the source. On the other hand, Vehicle2b, having its connections inversely, will behave different Figure 2. Nothing will change if the sensors are excited same unless we change sensor functions, will be touched on later. But consider again right sensor senses less light, the light source is at left side.
This time left motor will work less, and Vehicle2b will turn to light source. We can think as it will recognize light source from wherever it is, and run towards it, and acceleretaes, and BOOM!
Braitenberg defines these vehicles based on their behaviours; Vehicle2a as coward , since it dislike the source and escapes when it sees, and Vehicle2b as aggressive , since it tracks and tries to destroy the source no matters what.
Now having felt a little bit the wind of vehicle emotions, we can investigate how to vary them. Vehicle3a and Vehicle3b will be our next guests. As I mentioned earlier, we can change the sensor-motor functions. The simplest way to do that is to reverse, in other words, now our vehicles will move slower when they sense more. So we keep connections same as they were in 2a and 2b, but make sensors have an inhibitory effect on motors Figure 3. Thus, Vehicle3a and 3b will decelerate without changing their orientation when their both sensors are excited the same.
Figure 4 Moreover, there can be conditions where they face the light source and just stop. Analyzing Vehicle3a, it detects light source, faces it then decelerates as it gets near.
Then stops and watches the light source. Whereas Vehicle 3b accelerates when it can't see the light source, and when it finds, it just decreases speed.
Yes, Vehicle3 types both like the source, but they show their love in alternating ways. Vehicle3a loves the source in a permenant way, that it can't take its eyes from the light. On the other hand, Vehicle3b likes the nearby light allright, but it is always ready for another more powerful one, looking for it.
That's why Braitenberg calls Vehicle3b as an explorer. But this time we will come up with a third model, Vehicle3c. Think of all the four type of vehicles that we encountered up to now.
We will combine them in Vehicle3c. Vehicle3c has eight sensors, two are Vehicle2a's sensors - uncrossed and excitory, two are Vehicle2b's sensors - crossed and excitory, two are Vehicle3a's sensors - uncrossed and inhibitory and last two are Vehicle3b's sensors - crossed and inhibitory Figure 4.
But of course these all eight sensors should not sense the same environment variable. In pairs, they will be reactant to light as usual , temperature as our first model , oxygen concentration and organic matter concentration.
Finally we have a vehicle, robot or creature; that escapes from hot places but always stays at places full of oxygen, tries to destroy light bulbs and likes organic matters. Looking it as a whole, Vehicle3c has a system of values , and has knowledge that it tries to destroy light bulbs because it knows they heat the environment, or knows chemosynthesis because it prefers places full of oxygen and organic matter. As we are aware of the possibility to change sensor functions, henceforth we will play with them.
Up to know all vehicles had direct or inverse proportionbetween its velocity and the amount of environment variable. In vehicle4 brand of robots, the graph of the function is like a hill, having a maximum speed in avarage light, decelerating when light changes positively or negatively. We can imagine that from dark to avarage light, Vehicle4a acts like Vehicle2a then from avarage to maximum light, it acts like Vehicle3a.
That causes the vehicle to orbit around the source or drawing eight figures or some more different shapes Figure 5a - Figure 5b. As the tastes of Vehicle4as with different functions are too varied, this complicated type of behaviour may make us think that these creatures have instincts.
But we can make it even better than just having instincts. Another possible change in functions is to put a threshold. The vehicle will not be interested in the source untill it passes the threshold value, which makes the vehicle seem like it make decisions about the environment. And taking this decision making system as a basis, Braitenberg concludes with the traces of existance of free will in his vehicles. Then making progress in this threshold issue, it is possible to combine some threshold devices between motors and sensors or make complex input devices to create some logic instances.
These threshold device complexes will be used to make vehicles' brains from now on. For example, it can know other robots, it can count, it can visit every tenth source on its way, or even it can escape from vehicles which have sensor number multiple of seven, believing that they bring bad luck. Figure 5a Redounding the number of sensors and connections, it is also possible to make a puzzle-solving vehicle!
But there is a gap, for these more complicated calculations a memory is needed, and the solution is still there. It is possible to make memory with these threshold devices. Take two threshold devices, the first is connected to a red light sensor consider as input and the other is connected to the first one. Once a red light activates the first threshold device then it activates the second and the two devices will activate each forever, which shows that they have been activated once upon a time.
This storage capability is the main block of a memory which can then be used to form real memories, much more complex but finite in the number of threshold devices.
Even this finity situation can be solved. Think of a vehicle capable of leaving marks at the environment. It can not comprehend the marks but read and measure them to give a result. Then this vehicle can do almost everything that a computer does, in an infinite area that it can leave marks on. The challange for vehicles is not making more and more complex things, but to survive. Their toy world deserves to be resembled to real world especially with the war for survival in addition to environment and creatures.
Like all species in the world, vehicles are also subject to natural selection, modification and evolution as well. As gods of the vehicle world, we experiment this evolution like this: First we release all brands of vehicles on a table, designing the environment to include light, temperature, oxygen, smell, sound, undulating land, everything that can affect the evolution.
Lab 5: Braitenberg Vehicles
Not for the amusement: science fiction in the service of science. Or just science, if your agree that fiction is a part of it, always was, and always will be as long as our brains are only minuscule fragments of the universe, much too small to hold all the facts of the world, but not too idle to speculate about them. Valentino Braitenberg. This project aims to simulate some primitive robots which are given the name "Braitenberg Vehicles" , after first being stated in Vehicles: Experiments in Synthetic Psycology. Before giving information about the simulator and the project, it is better to have some idea about the book and its enthralling author, Valentino Braitenberg. Valentino Braitenberg was born in , Bolzano. Being a German neuro-scientist and cybernetics, he is a specialist in communication and movement control.
Some of Valentino Braitenberg's most well-known work centers on thought experiments with what he calls "vehicles. Each experiment details a vehicle that has a small set of sensors, and how those sensors can be connected to the vehicle's motors in ways that mirror the neurological connections in living creatures. The resulting vehicles seem to be capable of complex behaviors like fear, aggression, love, free will, etc. In his work, Braitenberg describes what he calls the Law of Uphill Analysis and Downhill Invention: "It is much more difficult to guess the internal structure of an entity just by observing its behavior than it is to actually create the structure that leads to that behavior. In this lab, you will program your robot to emulate five of Braitenberg's vehicles, using the light sensors as the "small set of sensors" Braitenberg describes. You will begin with the most basic vehicle, Alive.
An Introduction to the Analysis of Braitenberg Vehicles 2 and 3 Using Phase Plane Portrait
Introduction When looking at mechanisms with cognitive functionality and artificial intelligence in general it is useful to begin with the simplest cases. And yet we know very well that there is nothing in these vehicles that we have not put in ourselves. The speed of the motor rectangular box at the tail end is controlled by a sensor half circle on a stalk, at the front end. Motion is always forward, in the direction of the arrow, except for perturbations. Vehicle 2a: fear pp. If the stimulation is directly ahead, the vehicle may hit the source.