AnimatLab is a tool that combines biomechanical simulation and biologically realistic neural networks. You can build the body of an animal, robot, or other machine and place it in a virtual world where the physics of its interaction with the environment are accurate and realistic. You can then design a nervous system that controls the behavior of the body in the environment.
The software currently has support for simple firing rate and leaky integrate and fire spiking neural models. In addition, there a number of different synapse model types that can be used to connect the various neural models to produce your nervous system.
On the biomechanics side there is support for a variety of different rigid body types, including custom meshes that can be made to match skeletal structures exactly. The biomechanics system also has hill-based muscle and muscle spindle models. These muscle models allow the nervous system to produce movements around joints. In addition, there are also motorized joints for those interested in controlling robots or other biomimetic machines. This allows the user to generate incredibly complicated artificial lifeforms, otherwise known as animats, that are based on real biological systems.
Best of all, standard AnimatLab is completely free and includes free source code!
If you'd like to, you can upgrade the standard edition to the professional one. The pro version is for users who want to to do some real research with AnimatLab or who want to build more complex simulations. This Pro edition has to be purchased.
v2.1 [Apr 3, 2015]
1. You can now add robot part interfaces to your existing simulations to control real hardware using the neural network control system from your simulation. Please see the Robotics documentation for more details.
2. AnimatLab now has a framework in place to allow physical devices to interact with a neural network in a simulation or on a robot. Right now only the wireless XBee commander joystick is supported, but I plan to add keyboard and standard joystick interactivity soon. Please see the remote control documentation for more details.
3. I have created realistic mesh files for all parts of the base Bioloid robot frame kit. They are available for free in the AnimatWarehouse.
4. Contact sensors can now become "sticky." Your neural network can turn on/off the stickiness of a contact sensor, and when it is on and comes into contact with another part in the scene they will stick together until stickiness is turned off. I used this in the robot arm simulation to allow the gripper to pick up parts. However, there are a number of other applications. For example, you could simulate a gecko's ability to walk up walls.
5. You can now specify a target data type for adapters. Previously the incoming data type for a node was hard-coded. Now it is possible for the user to change it. The main place where this is used is to allow you to have two adapters connected to a joint where one lets you control the target position and the other sets the target velocity.
6. You can now setup a delay buffer on any adapter, and you have the ability to disable all output from an adapter for a given period immediately after the simulation starts in order to allow the neurons in your simulation to settle down to a steady state.
7. Added a Modulate Neuron Property synapse to the Firing rate neural plug-in. This allows you to modify any of the properties of the post-synaptic neuron based on the firing rate of the pre-synaptic neuron.
