bam.model#

Attributes#

models

Classes#

Model

Friction model for a servo actuator.

DummyModel

Friction model for a servo actuator.

Functions#

load_model(→ Model)

Load a BAM friction model from a parameter file.

load_model_from_dict(→ Model)

Load a BAM friction model from a parameter dictionary.

Module Contents#

class bam.model.Model(load_dependent: bool = False, directional: bool = False, stribeck: bool = False, quadratic: bool = False, name: str = None, title: str = '')#

Friction model for a servo actuator.

Combines a motor model (set via set_actuator()) with a friction budget that maps joint state and torques to a maximum resistive torque \(\tau_{fm}\). The friction applied by the simulator is the stopping torque clipped in \([-\tau_{fm}, \tau_{fm}]\).

Use load_model() to instantiate from a parameter file rather than constructing directly.

Parameters:

  • load_dependent – Enable load-dependent friction terms.

  • directional – Enable directional (motor-side vs external-side) load friction. Requires load_dependent=True.

  • stribeck – Enable Stribeck effect (higher friction near zero velocity).

  • quadratic – Enable quadratic load-friction coupling term. Requires directional=True and stribeck=True.

  • name – Model variant identifier, e.g. "m6".

  • title – Human-readable title used in plots.

actuator_name = None#
name = None#
title = ''#
load_dependent: bool = False#
directional: bool = False#
stribeck: bool = False#
quadratic: bool = False#
max_friction_base = 0.2#
max_load_friction = 0.5#
max_viscous_friction = 1.0#
reset() → None#

Resets the model internal state

set_actuator(actuator: bam.actuator.Actuator) → None#

Attach an actuator to this model and initialize its parameters.

Parameters:

actuator – Actuator instance to attach. Determines the motor model (kt, R, kp, vin) and the set of friction parameters that are created on this model.

compute_frictions(motor_torque: float, external_torque: float, dtheta: float) → tuple#

Compute the friction budget for the current state.

Returns the two MuJoCo friction parameters that implement the BAM friction-budget formulation: frictionloss (Coulomb-like constant term) and damping (viscous term). The simulator applies friction by clipping the stopping torque in \([-\text{frictionloss} - \text{damping}\cdot|\dot{\theta}|,\ +\ldots]\).

Parameters:

  • motor_torque – Torque produced by the motor [Nm].

  • external_torque – External (gravity/load) torque seen at the joint [Nm].

  • dtheta – Joint velocity [rad/s].

Returns:

Tuple (frictionloss, damping) ready to be written into mj_model.dof_frictionloss and mj_model.dof_damping.

get_parameters() → dict#

This returns the list of parameters that can be optimized.

get_parameter_values() → dict#

Return a dict containing parameter values

load_parameters(json_file: str) → list#

Load parameters from a given filename

load_parameters_from_dict(data: dict) → list#

Load parameters from a given dict

class bam.model.DummyModel#

Bases: Model

Friction model for a servo actuator.

Combines a motor model (set via set_actuator()) with a friction budget that maps joint state and torques to a maximum resistive torque \(\tau_{fm}\). The friction applied by the simulator is the stopping torque clipped in \([-\tau_{fm}, \tau_{fm}]\).

Use load_model() to instantiate from a parameter file rather than constructing directly.

Parameters:

  • load_dependent – Enable load-dependent friction terms.

  • directional – Enable directional (motor-side vs external-side) load friction. Requires load_dependent=True.

  • stribeck – Enable Stribeck effect (higher friction near zero velocity).

  • quadratic – Enable quadratic load-friction coupling term. Requires directional=True and stribeck=True.

  • name – Model variant identifier, e.g. "m6".

  • title – Human-readable title used in plots.

bam.model.models#
bam.model.load_model(json_file: str = None, *, motor_name: str = None, model: str = None) → Model#

Load a BAM friction model from a parameter file.

Specify the source with one of two mutually exclusive approaches:

  • Bundled motor — pass motor_name and model:

    model = load_model(motor_name="xl330", model="m6")

  • Custom JSON — pass json_file (output of bam.fit):

    model = load_model("params/my_motor/m6.json")
Parameters:

  • json_file – Path to a BAM params JSON file.

  • motor_name – Name of a bundled motor (e.g. "xl330", "mx106"). Must be combined with model.

  • model – Model variant for a bundled motor ("m1"–"m6"). Must be combined with motor_name.

Returns:

A Model instance with all parameters loaded.

Raises:

FileNotFoundError – If the requested bundled motor or model does not exist.

bam.model.load_model_from_dict(data: dict) → Model#

Load a BAM friction model from a parameter dictionary.

This is the low-level counterpart of load_model(), used when the JSON data has already been parsed.

Parameters:

data – Dictionary produced by json.load on a BAM params file. Must contain at least "model" and "actuator" keys.

Returns:

A Model instance with all parameters loaded.