Accelerometer#

The accelerometer sensor is used to measure the three-axis linear acceleration at the mounting point in the local coordinate system. In robot simulation, accelerometers are commonly used to perceive the robot’s motion state and provide important feedback information for control algorithms.

🎯 Functional Description#

The accelerometer measures the linear acceleration at the sensor mounting point in the site local coordinate system, including the influence of gravitational acceleration. This sensor returns an array of 3 floating-point numbers, representing the acceleration components in the X, Y, and Z axis directions respectively.

📋 Return Value Format#

acceleration = model.get_sensor_value("accelerometer_name", data)
# Type: numpy.ndarray[float32]
# Shape: shape = (*data.shape, 3)
# Unit: m/s²

  • acceleration[…, 0]: X-axis acceleration component (local coordinate system)

  • acceleration[…, 1]: Y-axis acceleration component (local coordinate system)

  • acceleration[…, 2]: Z-axis acceleration component (local coordinate system)

⚙️ MJCF Configuration Parameters#

In MotrixSim, accelerometer sensors support the following MJCF configuration fields:

Basic Configuration#

<sensor>
    <accelerometer name="sensor_name"
                   site="site_name"/>
</sensor>

Supported Attributes#

Attribute Name

Type

Required

Default

Description

name

string

✅

-

Unique identifier name of sensor

site

string

✅

-

Name of reference point where sensor is mounted

Note: MotrixSim currently does not support the cutoff, noise, and user attributes from the MJCF standard.

📝 Configuration Examples#

Basic Accelerometer Configuration#

<!-- Define mounting point in body -->
<site name="car_imu" type="sphere" size="0.03" rgba="0 1 0 1" pos="0 0 0"/>

<!-- Define accelerometer sensor -->
<sensor>
    <accelerometer name="car_accelerometer" site="car_imu"/>
</sensor>

Multiple Accelerometer Configuration#

<!-- Install multiple accelerometers at different positions -->
<site name="imu_base" pos="0 0 0" size="0.02"/>
<site name="imu_end" pos="0 0 1" size="0.02"/>

<sensor>
    <accelerometer name="acc_base" site="imu_base"/>
    <accelerometer name="acc_end" site="imu_end"/>
</sensor>

🚀 Usage Examples#

Python API Usage#

import numpy as np
from motrixsim import load_model, SceneData, step

# Load scene
model = load_model("scene_with_accelerometer.xml")
data = SceneData(model)

# Run simulation and get accelerometer data
for step_count in range(1000):
    step(model, data)

    # Get accelerometer data
    acc_data = model.get_sensor_value("car_accelerometer", data)

    # If single environment simulation, data shape is (3,)
    if acc_data.ndim == 1:
        print(f"Acceleration: [{acc_data[0]:.3f}, {acc_data[1]:.3f}, {acc_data[2]:.3f}] m/s²")

        # Calculate acceleration magnitude
        acc_magnitude = np.linalg.norm(acc_data)
        print(f"Acceleration magnitude: {acc_magnitude:.3f} m/s²")
    else:
        # Vectorized environment case
        print(f"Acceleration data shape: {acc_data.shape}")

Practical Application Scenarios#

# Robot fall detection
def detect_fall(acceleration, threshold=15.0):
    """Detect fall based on accelerometer data"""
    acc_magnitude = np.linalg.norm(acceleration)
    return acc_magnitude > threshold

# Stationary detection (considering gravity)
def is_stationary(acceleration, gravity_threshold=9.5):
    """Detect if device is stationary (considering gravity influence)"""
    acc_magnitude = np.linalg.norm(acceleration)
    # When stationary, acceleration magnitude should be close to gravitational acceleration
    return abs(acc_magnitude - gravity_threshold) < 0.5

# Calculate tilt angle
def calculate_tilt_angle(acceleration):
    """Calculate device tilt angle based on accelerometer"""
    # Assume device is mainly affected by gravity
    gravity = np.array([0, 0, -9.81])

    # Normalize vectors
    acc_norm = acceleration / np.linalg.norm(acceleration)
    gravity_norm = gravity / np.linalg.norm(gravity)

    # Calculate angle
    cos_angle = np.dot(acc_norm, gravity_norm)
    angle = np.arccos(np.clip(cos_angle, -1.0, 1.0))

    return np.degrees(angle)

📊 Physical Principles#

The accelerometer measurement is based on Newton’s second law. In the simulation environment:

  1. Local coordinate system measurement: The returned acceleration values are represented in the site’s local coordinate system

  2. Gravity influence: When the sensor is stationary relative to the ground, it will measure the projection of gravitational acceleration in the local coordinate system

  3. Linear acceleration: Measures the pure linear acceleration at the mounting point, excluding rotational components

For example, when a robot is tilted, even when stationary, the accelerometer will produce non-zero readings due to the projection of gravity in the local coordinate system.

⚠️ Important Notes#

  1. Local coordinate system: The returned acceleration values are represented in the site’s local coordinate system, not the global coordinate system

  2. Gravity influence: When stationary, the accelerometer will measure the projection of gravitational acceleration in the local coordinate system

  3. Mounting position: The sensor measures the acceleration at the site mounting point position

  4. No advanced attribute support: MotrixSim currently does not support cutoff, noise, and user attributes

  5. Data type: The return value is numpy.ndarray type, with shapes supporting vectorized environments