just need a simple idea on how does animal locomotion models its cyclical gaits and trajectories based on what, a mnimum amount of actuators thatis, weight, gravity, attitude and direction?
the appendages considered would be pic here, from front leg to hind leg and then every single body parts marked as the line
now to be fair, i have read most simulation research and all thesource codes, eliminates those that uses mocap data since i want something that is dynamic, not some roto materials and i gotta say, i just dont really understand those paper and how to, expand those models into other types of animals other than the passive and bipedal walkers, so yes you need to, go a little bit lower, and apply those 4 actuators into the appendages in picrel here. basicallh reexplaining the basic and how locomotion gait happens and their symmetrical pattern and its self balancing kinematics, and the following trajectory curve that comes. keep it short and direct and most of all,. basic.
thats all. thank you very much, and regardless of whatever this thread be, i have considered other options and is merely douvle checking this with /sci/ence and will continue on my own and posts here sparingly. thank you
Anonymous
>>14316974 >not using fore and aft, port and starbird Im am unable to decypher you nonsense, please seek medical assistance.
guessimjustgonnadomyownbloodyresearch
Quoted By:
>>14316993 i also dont know whats starbird so you need medical help too
Anonymous
>>14316993 LH/RH for hands? Hind? LFF/RF for foot/feet?
Sir, your diagram is all fucked up. I suggest you not pick more fights with ne and instead you should have pointed out my spelling error in the previous post.
>you not your >i also dont know whats starbird I know but its a standard orientation of direction from a centeral point. That like saying "I dont know what math is."
>Ok, I understand, let me show you. I didnt even read the post, just giving free bumps. I just could not deduce the specific meaning of it and thought I would help you. A proper question can get you tge proper answer.
Anonymous
Quoted By:
>>14317009 ...and then I respond to my own post...fucking embarrassing...
guessimjustgonnadomyownbloodyresearch
>>14317009 you noticed all those but stopped at tldr didnt you? this thread aint for you. nvm bumping
Anonymous
>>14317012 I literally only looked at the picture. Thats it. I dont know what this thread is about Im too autistic for this conversation, just interjecting autism.
guessimjustgonnadomyownbloodyresearch
Quoted By:
>>14317013 ok genius, sure whatever
guessimjustgonnadomyownbloodyresearch
Quoted By:
im done here. everyone s too advanced already for a sound communication
Anonymous
>>14316974 Oh hey it's you again. What exactly are you working on? What is this for?
guessimjustgonnadomyownbloodyresearch
>>14317105 >what is this for a sayonara
supposedly a simpler model of walking animation which only need you to draw a schematics in 3d space and actuators so one can know how the walking trajectory work instead of, building a whole robot
or running a program...
but i did remember i have asked "natural looking animation" so this may be the reason why everything now sounds off.
Anonymous
guessimjustgonnadomyownbloodyresearch
Anonymous
>>14317150 You wont be able to create an animation from an arbitrary skeleton, you will need to simulate physics and try different animations until you get closer and closer to something that looks natural for the skeleton you're simulating.
Anonymous
>>14317409 Don't bother, if OP can't find one of the 5183540396 results on google for "Reverse Kinematics" you won't be able to help him
guessimjustgonnadomyownbloodyresearch
>>14317409 what if simulation not available? was the passive walker not build from blueprints with gravity considered?
guessimjustgonnadomyownbloodyresearch
Quoted By:
>>14317413 explain how does "mathemathically figuring out the end position of a kinematic chains" actually describe an entire locomotion gait with at least 2(two) appendages, each having their own virtual connector that is inverse kinematic then?
Anonymous
>>14317413 Inverse kinematics won't be particularly useful for what OP wants. IK only calculates the joint rotations needed for the end effector to be at a particular position. How can you use that to generate a natural looking walk cycle animation for arbitrary skeletons?
>>14317419 Then you won't be able to generate the animation. An animation only looks natural if it obeys physics. Our brains are really good at understanding physics so when we look at an animation that seems to not obey physics it will seem unnatural and robotic.
Anonymous
You post this on /g/, right? It might be more their field.
guessimjustgonnadomyownbloodyresearch
>>14317431 >wont be useful cant say it wont be entirely. i am still beating around the bush on this topic.
>unable to then how does robotics model natural motion without using mocaps? it technically relied on its own model since a mocap wont match it realistically
>good at understanding true, we can tell if an animation is wrong by memory, but the purpose of this is to, create a logical one, relying on the blueprint i have set, like the above. granted i ve not seen ALL scientific invention on robotics, but a particular simulation is said to "not use motion capture data" to perform natural looking gaits, thus means relying on its own actuators, properties and mechanism. the paper was also quite complex and only support bipedal walker... maybe i can find it again later but generally, the model i require would have been a sort of general blueprint of how everysingle terrestrial locomotion walk themselves.
Anonymous
>>14317451 Nah, /g/ is too busy jacking off to their consumerist electronics and installing gentoo.
I wish OP was able to write english properly as this is a really cool problem to discuss.
Like okay sure you can add physics to your skeleton but then it will just collapse to the ground.
How do you decide which muscles to activate to actually start walking after physics is applied?
It's a really difficult problem which is why it makes it interesting.
guessimjustgonnadomyownbloodyresearch
Quoted By:
>>14317451 programming codes not needed. passive walkers are /sci/s invention before computers, if i recall.
i d be on g for opengl stuff.
Anonymous
>>14317464 IK might be useful to determine which muscles to activate after you apply physics to your skeleton sure but you still need to simulate physics. There's no way around that. You first apply physics to the skeleton and then start activating muscles in ways that makes it walk. Initially its just going to flop around on the ground like a dead fish not moving, but eventually if you design it right it will learn to walk.
Anonymous
Quoted By:
>>14317465 >Nah, /g/ is too busy jacking off to their consumerist electronics and installing gentoo. Christ thats sad.....
guessimjustgonnadomyownbloodyresearch
>>14317465 well pardon for the english since i am not exactly born and bred white man.
logically the muscles are always activated, for this particular walking case, or even any sort of crawling, as for the difficulties i thought this has been solved and iterated to various other subjects??
the passive walker has been around since 1980s
but then animations didnt have those in 1959 either...
guessimjustgonnadomyownbloodyresearch
>>14317484 technically speaking all the muscles are supposed to be"activated", while considering their attitudes, limping, crawling, running.
as for the "physics to the skeleton" i thought theres only weight and gravity? so then the only thing left is the biomechnical energy that "pushes" the body to walk which supposed to cycle in timely manner? is this not a /sci/ thing with mechanical walkers
Anonymous
>>14317492 >>14317499 By "muscle activation" i mean mimicking how real muscles work. You can only control the stiffness and force of each joint, the physics actually moves them. The stiffness of the joint tells physics how resistant to change the joint is, so if its very stiff it will take a lot of external force on the joint to move it, if its not stiff at all it will take very little force to move it. The force on the joint is the force that the joint applies externally to start changing its angle. If the joint is not pushing against anything it will require very little force to move it, but if the joint is pushing against something it will require a lot of force to move it. Keep in mind joints can push against other joints depending on how you've set up the skeleton, not just the floor or other objects in the world. You need to apply physics to everything in order to simulate it. It's not just weight and gravity, you need forces, momentum, stiffness, etc.
Anonymous
>>14317499 >as for the "physics to the skeleton" i thought theres only weight and gravity? And tendons? Like Achilles tendon, otherwise it would walk like it had a peg-leg pirate.
guessimjustgonnadomyownbloodyresearch
>>14317509 >mimicking how muscles work well that would be the ground blueprint for us to identify, with inverse kinematic considered, it becomes simpler to actuate muscles in a straightline from the ankle to the hip...
>joints pushing well the joints "lifts up" in a normal walk so it shouldnt push against anything so it is more of the question of when or why does it "lift up"
>joints pushing each other yes, this would be the orientations, also a case to consider but im not sure where to even start...
>momentum what is this? would be easier to see a schematic, i really thought biomechanism would have a study on this in a paper of sort
guessimjustgonnadomyownbloodyresearch
Quoted By:
>>14317529 >peg leg the idea is to just walk an arbitrary skeleton, so it wont look pegleg unles the design does so or mimic motion
Anonymous
>>14317532 >well that would be the ground blueprint for us to identify A blueprint is a technical drawing used in architecture and has nothing to do with this.
>to actuate muscles in a straightline You don't move the joints, you just apply forces on them and control their stiffness. The movement of joints is done by physics.
>when or why does it "lift up" When you jump you also "lift up" because you've applied an upward force to your body that propels you upwards until gravity overtakes it and you start falling down. Same with walking. You push from the ground giving the leg an upwards force which propels the leg upwards for a short amount of time. This has to all be controlled by physics, not by the animation.
>momentum: what is this Momentum is mass in motion. If something is moving it will continue moving unless an external force is acting on it because it has momentum.
>>14317529 You don't actually need to simulate actual muscles internally, you just need to control the force and the stiffness of each joint and let physics do the rest. You don't need tendons.
guessimjustgonnadomyownbloodyresearch
>>14317565 ok, this is gettin abit everywhere but this is the trajectory of a normal leg in walking animation,and given the properties of the character i simply need to know how is this trajectory came to be in relation to physics.
Anonymous
>>14317582 This trajectory will be different for different skeletons. You can't know what the trajectory will be in advance without simulating physics and it will change based on the angle of the floor you're walking on, objects you're walking across, current momentum and forces on each joint, etc. You can't control the trajectory, you can only control forces on the joints and stiffness and let physics move the skeleton based on those.
Anonymous
Fuck off.
To other /wsr/ anons. OP is a massive faggot that has been posting threads asking about gait and perspective for months across /wsr/, /g/, /sci/, /ic/, and perhaps other boards.
please ignore this thread. you'd be wasting your time trying to help.
OP typically has multiple threads active in any given catalog at once, such as this one
>>>/wsr/1176787 >>>/wsr/1176787 >>14316840 >>14316430 >>14316728 and to OP, pls go away
Anonymous
>>14317628 I don't care, it's an interesting problem.
Anonymous
>>14317667 Agreed. Its very fundamnetal and rudimentary, a"first pricipals" type problem. Finding solutions like these is similar to discoverying new physics/matys etc in that its "so obvious its hard to see".
More weird gait threads plx.
Anonymous
>>14317681 If it's so simple then go ahead, show us the equation for moving the joints of arbitrary skeletons such that it will look like a natural walk. Make it work for bipedal, tripedal, quadrupedal and n-pedal creatures. Not only that, make it work for randomly generated skeleton structures. Guess what, it's an extremely difficult problem that likely doesn't have an analytical solution so such an equation does not exist. It's interesting because its not easy.
guessimjustgonnadomyownbloodyresearch
>>14317608 yes it ll be different, so the key here to is to draw out the correct physic direction? i thought thats the basis of passive walkers and locomotion?
also i dont understand this stiffness of muscle since these arbitrary skeleton does not have muscles in application?
>>14317628 yea im kinda bored
>>14317667 >>14317681 >new invention actually im trying to research what exist...
guessimjustgonnadomyownbloodyresearch
>>14317699 despite all the robotics how come it does not exist?
Anonymous
Quoted By:
>>14317704 >actually im trying to research what exist I know. Have you ever teued to invent the concept of zero yourself? Every tried to calculate mass/time/enertia yourself, your own self made equations?
Same for this...if you are "doing your own research" you walking the path to its discovery.
Anonymous
>>14317704 Again, you can't calculate the trajectory in advance without simulating physics. You need to simulate physics to see the trajectory joints will take. Physics controls the trajectory, not you. You can only control the forces that joints exert, not the trajectory. Trajectory happens on its own and you can only observe it. A muscle in this context is the same as a joint.
>>14317708 Same reason why there is no analytical solution to the n-body problem. It likely does not exist. You can only calculate the path joints take by actually running the simulation and observing them. The difficult part comes from writing a program that decides what joints to apply forces to and when. For example, if you want to jump you need to first reduce the stiffness in your legs so your body starts falling down, and then apply a force to your legs very fast so the joint pushes away from the ground, gains momentum and starts flying upwards. The only thing you control is stiffness and forces, you don't control the motion itself. The motion is controlled by physics.
Anonymous
>>14317667 It IS an interesting problem, but OP is an attention starved schizo so don't expect any real discussion, like in the literal hundred of threads he created on multiple boards since a few months ago
guessimjustgonnadomyownbloodyresearch
>>14317721 how is physics simulated on paper then? like designing a ship, a car, they dont always had physics on the blueprint right? how do you "control" it?
>>14317726 with all your crying to the mod about your thread about your tiny mandible it s very hard to make any discussion
Anonymous
>>14317726 I know. I don't care. I will still reply.
>>14317747 Physics cannot be simulated on paper. You need to run a simulation to see what happens. You can run a simulation on paper yes and that's how NASA calculated trajectories before computers came along but that's very slow and time consuming. It is much faster to calculate trajectories using a computer simulation.
guessimjustgonnadomyownbloodyresearch
>>14317764 ok but this is not exactly rocket science so NASa is not necessary.
Anonymous
>>14317789 No it's not rocket science but its still a physics problem so you need to simulate physics, there's no way around that. There is no analytical solution. You need to simulate physics on your skeleton to see how well it walks when you change forces and stiffness of joints. And then you need to adjust your program that applies forces and stiffness to make it walk better and better until it has learned to walk naturally.
guessimjustgonnadomyownbloodyresearch
>>14317801 how was the passive walker designed so that it walks then?
and even if i simulate physics, the skeleton will just fall apart without a dynamic model, or mocap data, so you still need to write up the right walkig methoda and estimate poses otherwise you d just be simulating everything than actually knowing how it works , as this thread is about.
Anonymous
>>14317810 Passive walker is a robot with a single joint that either only walks downwards or uses a single motor to replace gravity. Very simple physics problems do have analytical solutions yes, but as soon as you add more joints and complexity you will need to run a simulation. In the case of a passive walker you just need to calculate the frequency needed for the joint to actuate and it will keep walking. However even for passive walkers it is much easier to just try out different frequencies until you find out the one that yields the best walking gait. So even in the case of a passive walker you are simulating physics until you find the right frequency.
Yes if you simulate physics the skeleton will just collapse to the ground which is why this is a difficult problem. You need to figure out a way to actuate the joint and control stiffness such that the skeleton doesn't collapse to the ground and instead starts walking. And no you can't use mocap data since you don't have mocap data for arbitrary skeletons, you need to figure out the best actuations from scratch by running the simulation and seeing what works and what doesn't work.
guessimjustgonnadomyownbloodyresearch
>>14317856 hmmm yes i think i have the simulation, a manual simulation but simulation nevertheless, but still this frequency will only work if the main actuations are decided like a schematics of sorts that will definite a correct motor pattern, so then it becomes a question on how this motor pattern came to be, which is done by guessing out trajectories? it s basically THE simulations.
Anonymous
>>14317878 This has nothing to do with any "schematics". If by "manual" simulation you mean running the same calculations as the computer does but on paper then sure theoretically you can do that but it's going to take you thousands of years to run a simulation. Physics simulations are calculated a fraction of a second at a time and you need to calculate physics for several seconds to see how well your motor pattern works for your skeleton. You can't calculate the correct motor pattern without simulating physics and trying different patterns out to see which one works best. You need to run the physics to see which motor pattern works best. You can't guess the trajectories, you need to run the simulation and observe them.
guessimjustgonnadomyownbloodyresearch
>>14317900 yes it' just probably a single second, which is 24 frames or so. Like a simple linear interpolations maybe.
I suppose it "exist" right?
guessimjustgonnadomyownbloodyresearch
>>14317900 as in it's just a series of coordinate in 3d space, so a lerp will do/is the correct model or maybe vector. Sorry i cant put it enough in english.
guessimjustgonnadomyownbloodyresearch
>>14318766 as in can you iterate a model of simulation that exactly matches this specific simulation's trajectories? Whatever thousand year system?
or maybe a straightforward model?
Anonymous
>>14316974 you are my favourite schizo on here mate
guessimjustgonnadomyownbloodyresearch
Quoted By:
>>14318930 and i hate your cry-alot
Anonymous
>>14317667 It is a fascinating problem, but OP is being a faggot of the highest degree about it and encouraging this kind of shit will only make the board even shittier in the long run. OP could have left any one of the last 50 threads up, instead every time he gets a few dozen responses he makes a new thread asking the same questions and ignoring *everything* that was discussed in the previous one.
At the time of posting this there's literally two more threads from this faggot active, another three that were *just* bumped into the archive within the last couple hours, and another half a dozen that got deleted by the mods.
>>14316430 >>14316840 >>14304845 >>14310537 >>14310867 All within the last 72 hours.
There's being passionate about a subject and then there's being a spamming faggot, and if you reward him for being an obnoxious shit then it just means the next time he has a homework project due or whatever the fuck this is, he'll run right back here and post 50 threads demanding the answer.
guessimjustgonnadomyownbloodyresearch
Quoted By:
>>14318994 point me what was discussed then, cryalot. let's see how smart you are.
guessimjustgonnadomyownbloodyresearch
>>14318994 meanwhile i m reporting you for your convenience. your 500 word cryalot essays can go to the mod because you cant afford to get your own cryalot containment thread, being a lazy faggot you are, you dont see you ever care about the the specifics at hand, this thread or any othr ones, but your cryalot essays and your worthless garbage english
Anonymous
>>14318745 Simulating physics has nothing to do with frames either. You can advance the simulation one microsecond at a time which means you will need to calculate physics 1000000 times per second which obviously won't be done in realtime. Simulating physics also has nothing to do with interpolation.
>>14318766 No, a lerp will not simulate physics. To simulate physics you need to know about joint forces, mass, rigidness, gravity, colliders, momentum, etc. Lerp itself is just a function that interpolates between two values. It's useless for physics.
>>14318994 Yes, he's retarded. No, i don't care. I will keep replying.
Anonymous
>>14319873 you know maybe you are thinking abit too far for this things i need. all i need is a simple trajectory/range in my walking animation which usually too is generated through interpolations with animation software.
Anonymous
>>14319056 >i m reporting you for your convenience Announcing reports is a bannable offense.
>>14317582 >>14318834 You have a final path. You want to find out the position of any legs and limbs as they trace this path. There are infinitely many positions the limbs can be in and still follow the path.
This is solved using inverse kinematics
https://youtu.be/hbgDqyy8bIw Anonymous
>>14320101 >you know maybe you are thinking abit too far for this things i need. You are a retarded person who spams huge numbers of technical words without understanding any of it. Your problems are incoherent because you do not know what you do not know.
Anonymous
>>14320101 Read
>>14317608 again. You can't generate the trajectory if you use physics. It's impossible. You need to simulate the physics and observe it. There's nothing to interpolate. Sure if you're manually animating the skeleton then yes you add keyframes and interpolate between keyframes but you don't want to manually animate the skeleton, you want to automatically generate the animation which is only possible by simulating physics.
Anonymous
>>14320107 No, this is not going to be solved with IK. Read
>>14317431 and
>>14317484 . What OP wants is to generate a natural looking animation for arbitrary skeletons, not move the skeleton so that the end effector is in a particular position.
Anonymous
>>14320107 >>14320114 >>14320117 if you cant do a simple lerp why should i take your word? even the computer does the interpolation cheaply, so it is not impossible to at least guess trajectories.
you are overcomplicating this matter.
>>14320114 i am not sure you are understanding me completely either becaue our idea is mismatched
>>14320120 correct.
Anonymous
>>14320120 >What OP wants is to generate a natural looking animation for arbitrary skeletons OP might as well be asking for a unicorn.
The amount of computational complexity he's asking for is ridiculous. You cannot have something that is both generic enough to determine natural motion for any arbitrary body and simple enough to actually employ.
Anonymous
>>14320219 Lerp (linear interpolation) is a process of finding a value between two values. So if you have a value 10 and value 20, a 50% interpolation between them would be a value of 15. An animation keyframe is the position of all joints in the skeleton at a particular moment of time. An animation is a collection of keyframes. If you are manually creating the animation you are creating those keyframes and saving them to a file. To play an animation, you can lerp between keyframes so that the animation is smooth. If you didn't lerp between keyframes, the animation would jump from one keyframe to the other which would not look smooth. That's why lerp is used.
You can't use lerp because you don't have any keyframes to work with. If you had keyframes you could lerp between them but you don't have any. If you want keyframes, you need to manually create them using an animation software. But you said are trying to create a natural looking animation from arbitrary skeletons automatically without manually animating it.
>>14320243 Sure but it's not impossible. You can apply physics to your skeleton and then start applying forces to joints to try to make it walk. The software you write has to try different ways of applying forces to different joints until it learns how to get it to walk and move. There's papers that have done exactly this already. It's not impossible it's just really really hard. Which is why it's interesting. Which is why i don't care that OP is retarded.
Anonymous
>>14320267 OP refuses to learn how to apply dynamics to a single rigid body but rest assured he's taking the problem super seriously
Anonymous
Quoted By:
>>14320243 >unicorn yeah but i didn.
>>14320267 >about lerp ok sure, seems i got it slightly wrong.
>animation without animation did i? the problem is that a natural looking animation needs correct natural trajectories, an arc they say so a base logic should be set first less so to build a logically natural keyframes throughout the scne.
>>14320814 >applying dynamics did i? i dont even know where to start.
it's as simple as defining gravity in vector 3 as 0,0,9.8 in xyz isnt it? and then the body just falls down, technically.