Hi, Sir. I have some questions about modeling program.
I want to build an ingress motion optimization model, but I don’t know that AMM can achieve it.
I want to build a car driver’s space interface by CAD, and simulate the specific ingress/egress strategy by adjusting the space design parameter, such as the door height or degree of A-pillar, because I want to know if the space design parameters change, what would be muscle (lower limb, etc.) activation varied? higher or lower?
Can I use mocap data to get the specific motion strategy, and auto-simulate the same motion strategy with change of space design parameters? Because the muscle activations related to ingress/egress motions that would be changed with change of space parameters, and the motions would have some difference (joint angle, etc.) with my original mocap data, but strategy is same. I hope you know what I mean.
The goal of the model is that I want to reduce the trials to build the similar model, use model to know difference of muscle activations by adjusting a small margin change of space parameters.
Can the Anybody do this? such as auto-simulation with adjusting some parameters (the motion strategy is same), and identify the good parameters with lower muscle activation.
It is a good question and I understand what you aim to do. We are beginning a research project about that exact topic within a couple of months from now.
So far, we have a few examples of optimization of motion, but for something as complex as ingress/egress it will be challenging. Here's the procedure:
[ol]
[li]Use a mocap experiment to generate a baseline model[/li][li]From this model, dump all the joint angle variations from the interface secton of the model (where the anatomical angles can be found)[/li][li]Downsample the angle variations to be interpolated between a few points, for instance 5 for each driver.[/li][li]Use these angles as input to the drivers from the FreePostureMove model to reproduc the motion.[/li][li]Formulate linear measures to calculate distances to obstacles in the model.[/li][li]Create an optimization problem with the downsampled interpolation points as variables and non-collision with the obstacles as constraints.[/li][/ol]You can see the technique in the tutorial Posture and Movement Prediction.
