Why such device doesn’t exist?
History
To understand the problems of building a walker for paralyzed people, one should first know the balance problems of two legged walking robots.
Basically today, there are two kinds of different walking systems for two legged robots. The older system named “Static Walking” and the more recent system named “Dynamic”.
If we ignore the energy usage; because the recent studies on batteries shows that it will not be a problem in the future; the main difference of dynamic walking from the static for the consumer is the walking speed. For making it more clear one can say that after the researchers solved the balance problem of two legged robots precisely, they tried to make the robot walk faster. But while they are doing that, they realized that even though the calculations are correct, the robots unexpectedly started to fall. That was the dynamic effect of walking.
To explain that effect first I will explain static walking. As anybody can easily figure out; for after like 5 minutes of thinking, if a two legged robot wants to lift one of his legs up, it has to bend its upper body to the other leg for not to fall. Continue from there for moving its leg to forward it has to bend his body backwards. The purpose of these actions is for keeping the center of mass (the projection of the center of mass on the ground) inside of the carrying foot.
 |
Old System |
 |
Although this is the more precise way of securing the balance, this walking system has two main negative sides. First of all this is too many moves for a single step. Bend the upper body in this order to first side and backwards, than other side and forwards, than other side and backwards, finally first side and forwards. Then the walking was too slow; it takes 20-40 seconds to make a single 15 cm long step.
It was that slow at the time that the scientists were using static walking system on robots. The problem here was; if you order the robot to bend its upper body to a direction too fast, it will definitely fall because of the momentum even though the projection of the center of mass (COM) remains on the right place. The foot of the robot should be at some other point. The name of this point is Zero Momentum Point (ZMP) and means, the point where the sum of all moments is zero.
The main essence of dynamic walking is to calculate the ZMP, but the problem is; they do not have the right formula yet.
To explain the difficulties of calculating the ZMP first I have to explain the balancing principle of dynamic walking. The dynamic walking starts with bending the robot to forward; until the projection of the COM falls more forward than the forward end of the feet. The robot begins to fall. Normally this fall accelerates with every passing millisecond. The robot has less than one second to move its leg forward and place it to the right spot. This spot is not the ZMP, because if it places the foot on ZMP than the whole system will stop moving. But the dynamic walking is a continues move. So it places the foot on the right spot, and the right spot is just a little backwards of the ZMP. It is the point which acceleration of the fall becomes zero but the fall continues. The robot has less than one second to place its other foot to the next spot.
Actually after calculating the ZMP it is not a big deal to calculate the right spot. But calculating the ZMP is impossibly hard; because everything affects the ZMP. The type of floor affects the ZMP, the angle of floor affects the ZMP, the softness of floor affects the ZMP, the slipperiness of floor affects the ZMP. Even the wind affects the ZMP. If there are some other people or objects in the room and the robot has to make sudden turns, it also affects ZMP. If the robot carries something, it also affects the ZMP. If there are any kinds of faulty on the floor, like a stone or a piece of trash or it has been three years since the pavement road has built and it starts to deform, it affects the ZMP. And moreover these calculations should be done in less then one second so this kind of advanced technology does not exist and won’t be able to reach this level of complicity during the next 10 years. Analyzing the floor before the step seems also impossible because not only the deformation but also the softness (like the softness of a carpet) of the floor should be detected by sensors.
It is a fact that to calculate the ZMP doesn’t seem so difficult for a stranger to this topic. As an example if you would search the internet there are lots of calculated realistic simulations about walking two legged robots. But with closer look at those, you would find out, almost all of those simulations are two dimensional. The difference between the calculation of two and three dimensional balance is; in three dimensional the robot can not only fall to front or back, but also can fall laterally; right or left side. And in real life robots always fall to sides, because this is not a kind of problem which can be solved by making the feet larger.
So one can ask how are the robots on television walking so fast and like a human today. The answer is they simply do not. For making their robots walk, the companies are building every object on that area special for the robot, starting with the floor. The steps are pre-calculated. They are so far away from producing their robots. If one is interested one can find that the robots 12 years before were walking as good as the ones that walking today. On that passed 12 years no robots are manufactured and multiplied for selling. No company now says that they will do it any time soon.
Why Current Walking Systems Are Not Suitable
After explaining the weaknesses of known walking system it is easier to explain that those systems are not suitable to produce such a device.
First of all it should be realized that, that kind of a suit is a motorized walking system or in other words the lower body of a walking two legged robot. Wearing the suit means to bind the patients’ legs with some kind of fabrics to the robots legs and placing safely the patient’s upper body to the upper side of the walking system. Then the system will also be covert to look better.
There are many reasons why it is impossible to make that kind of a walking suit with current balancing systems, but we can gather them in four main subtitles.
Because the suit is a walking system which wore by a paraplegic patient, it will not be able to do things like walking on surfaces other than perfectly flat. It can not climb any other stairs than the ones which designed for it. And the places of the objects at home will be very important for not to crumble. It will also not walk anyway if it is not pre-programmed. As an example if it not pre-programmed they could not decide to go nearby the table and take the glass because if the table is 25 cm away and programmed steps are 40 cm.
This reason by itself effects and reduces almost all of the market. But we are sure that there will be some (probably very rich) patients which will buy this product and will want to rebuild some parts of their homes. Paralyze people are ready to change major parts of their lives anyway. If this would be the only problem, many companies would already produce it and it would be a great selling, successful product.
This is unfortunately impossible. As it is mentioned before for the calculation of the ZMP, the upper body and its acts are extremely important. Even for five steps, it is impossible for a human to sway his arms with exact speed, acceleration, start - finish point and force. On the other hand this is a very easy thing to do for a robot. .
Today’s robots can not sense and calculate the surface of the floor. But to sense and calculate both floor and the body it carries at the same time, is multiple times harder than that. It should also be realized that the floor is a static body and it is much more possible to calculate it alone. But the patient’s body is dynamic and both sensing, calculating and making immediate new actions to the reaction of patient’s body to the actions is today impossible.
But I am an optimistic person, so I think in one decade or so it will be possible. Not the new generations of known systems but a whole new calculation principle will solve this problem. Then there is the third problem.
There is no possible way, even in theoretically, to know a human’s moves before he does them, other than putting some microchips into his brain. As it is mentioned before, making any wrong sway with hand or one degree of false wearing of this suit will affect the ZMP. The body will start to fall. As an example it starts to fall to the patients left, there is no other way than to take a quick step to the left to prevent falling. This quick step will shake the patient and because of medical reasons, like weaken muscles this quick step will cause edema. Edema is a major problem in paralyze patients lives and to prevent edema which is caused by immediate impacts has no known ways yet.
Those facts mentioned above are the reasons why to make this kind of a suit impossible for one particular patient. But to produce a device and to make it suitable for each patient with all different body types and problems is a different thing. All of those patients have different height and different weight. Most of them have calcification on their knee and belt joints. So even for straight walking there should be as many flow diagrams as the number of the costumers.
So there is no such product. After analyzing the known balancing
technologies it is not too pretentious to claim
that this suit will not be produced with known systems for many years.