ExoStack / ROS 2 interfaces

External inputs

Your own nodes: estimators and sensors that feed ExoStack, and drivers that take its set-points.

ExoStack accepts signals from your own nodes and hands its set-points to any ROS 2 driver. Publish on the topic named in the YAML file, with the type and layout shown, and use best-effort QoS to match ExoStack.

Topic	Description	Message type
`/<speed_topic>`	walking speed for the lookup table → profile node	`std_msgs/Float32`estimator input layout
`/<intent_topic>`	locomotion intent → profile node	`std_msgs/Float32`estimator input layout
`/<gait_phase_topic>`	gait phase → planner	`std_msgs/Float32`estimator input layout
`/<external_sensors_topic>`	metabolic, EMG or other signals → cost evaluator	`std_msgs/Float32MultiArray`external sensors layout
`/<additional_gait_marker_topic>`	in-cycle marker → cost evaluator	`std_msgs/Int32`gait marker layout

Driving another actuator

A planner's output_topic is an ordinary std_msgs/Float32 set-point, so a ROS 2 driver of any actuator can subscribe to it instead of an actuation hub.

A minimal estimator

speed_estimator.py

import rclpy
from rclpy.node import Node
from rclpy.qos import QoSProfile, QoSReliabilityPolicy
from std_msgs.msg import Float32


class SpeedEstimator(Node):
    def __init__(self):
        super().__init__("speed_estimator")
        qos = QoSProfile(depth=10, reliability=QoSReliabilityPolicy.BEST_EFFORT)
        self.pub = self.create_publisher(Float32, "/walking_speed", qos)   # speed_topic in profile.yaml
        self.create_timer(0.1, self.tick)                                  # 10 Hz

    def tick(self):
        self.pub.publish(Float32(data=3.5))                                # your estimate here


rclpy.init()
rclpy.spin(SpeedEstimator())