Steering Rear Axle
Steering Rear Axle – definition
Steering Rear Axle is a vehicle chassis configuration in which the rear axle, or the rear wheel pair, can change steering angle relative to the vehicle body. In mobile robotics, this term describes a UGV whose rear wheels are not fixed in heading and participate in steering, either alone or together with the front axle. The concept is used in kinematic modeling, control design, path tracking, and mechanical layout analysis.
In a standard Ackermann vehicle, steering is usually implemented on the front axle only. A steering rear axle extends this model by adding one more steering degree of freedom. Depending on the design, the rear axle may steer in the opposite phase to the front axle for a smaller turning radius, or in the same phase for crab motion and improved lateral path alignment. In robotics literature, this is often described as four-wheel steering, dual-axle steering, or coordinated steering. The exact term depends on whether both axles steer symmetrically and whether the wheel geometry follows Ackermann constraints.
For UGVs, a steering rear axle changes the feasible motion set of the platform. It affects minimum turning radius, yaw response, tire slip, odometry model, and controller structure. It also changes how velocity commands should be interpreted in ROS 2. A chassis with rear steering cannot be modeled correctly as a differential drive robot, and in many cases it also differs from a simple front-steered bicycle model.
Kinematic meaning in mobile robotics
The technical meaning of a steering rear axle is best understood through planar vehicle kinematics. If the front axle steering angle is denoted by δf and the rear axle steering angle by δr, the curvature of the vehicle path depends on both angles and on wheelbase L. Under a low-slip kinematic approximation, a common equivalent bicycle relation is:
kappa = (tan(delta_f) - tan(delta_r)) / L
where kappa is path curvature. This relation is simplified and assumes small lateral slip, rigid planar motion, and a suitable reference point. For opposite-phase steering, δr has opposite sign to δf, which increases curvature and reduces turning radius. For same-phase steering, front and rear axle headings become closer, curvature decreases, and the robot can move with reduced body yaw.
In practice, robotic platforms use several variants:
- Front axle steering only – rear axle fixed.
- Rear axle steering only – front axle fixed, less common in UGVs.
- Dual-axle opposite-phase steering – optimized for maneuverability.
- Dual-axle same-phase steering – used for crab steering and terrain alignment.
- Speed-dependent coordinated steering – rear steering gain changes with linear speed.
This distinction matters because localization, control, and simulation require the correct vehicle model. A planner or controller tuned for differential drive or skid steering will generally produce biased tracking on a rear-steered chassis.
How Steering Rear Axle is represented in ROS 2
ROS 2 does not define a single universal message type called “rear steering.” Instead, the concept is represented through robot description, controller interfaces, and command semantics. The baseline mobile command in ROS and ROS 2 is still geometry_msgs/Twist, but this message only expresses body-frame linear and angular velocity. It does not specify how those velocities are achieved mechanically.
For a robot with rear axle steering, the stack usually includes:
- a URDF model with steering joints for the rear wheels or rear steering linkage,
- a controller that maps body velocity targets to axle steering angles and wheel rotational speeds,
- odometry estimation that uses the correct chassis geometry,
- optionally
ros2_controlhardware interfaces for steering position and traction velocity.
In ROS 2, the robot description is typically published according to REP 103 for units and coordinate conventions, and frame naming follows REP 105 for mobile platforms. Steering joints are modeled in URDF with revolute joints, while wheel rotation may be represented separately. A coordinated steering controller then converts a target curvature into front and rear steering setpoints.
rear_steering_controller:
ros__parameters:
wheelbase: 0.82
front_track: 0.64
rear_track: 0.64
front_steering_ratio: 1.0
rear_steering_ratio: -0.5
max_front_steering_angle: 0.52
max_rear_steering_angle: 0.26
publish_rate: 50.0
The values above are illustrative. Actual limits must come from the platform specification, steering actuator range, and mechanical stop angles.
Key parameters and metrics
A rear-steered UGV must be characterized by geometry, actuator limits, and control bandwidth. These parameters determine whether the rear axle provides a practical mobility benefit or only adds complexity.
- Wheelbase
L– distance between front and rear axle centers, in meters. - Track width – left-to-right wheel separation on each axle.
- Rear steering angle range – for example, in radians or degrees.
- Steering rate limit – maximum angular change per second.
- Wheel speed limit – affects feasible curvature at a given velocity.
- Odometry update rate – often tens to hundreds of hertz in embedded ROS 2 systems.
- Command interface rate – often tens to hundreds of hertz depending on controller and bus.
- Minimum turning radius – derived from wheelbase and effective curvature.
The steering rate limit is especially important. If the rear axle can theoretically steer to a large angle but does so too slowly, the benefit is reduced during dynamic maneuvers. In autonomous navigation, this directly affects path tracking quality and planner feasibility.
Integration constraints on Leo Rover and Raph Rover
In Fictionlab platform context, Steering Rear Axle is mainly a chassis architecture concept rather than a default feature of Leo Rover. Leo Rover is a four-wheel skid-steer mobile robot. Its nominal drivetrain does not use front or rear steering axles. For that reason, the term should not be used to describe Leo Rover’s default motion model. Leo Rover is modeled as a skid-steer or differential-drive-like platform and supports ROS 2 through available community and vendor integrations, but ROS 2 support depends on the specific software image and integration path rather than being a universal native default.
On Leo Rover, rear steering would only appear in a custom engineering scenario. This would require a major mechanical redesign, additional steering actuators, updated URDF, a dedicated controller, and revised odometry. It would also invalidate assumptions used by standard differential-drive controllers.
Raph Rover is a larger platform with higher payload capacity. A steering rear axle is more relevant there in principle, because larger UGVs often benefit from reduced turning radius and improved maneuverability under load. If a Raph Rover configuration were equipped with coordinated steering on both axles, the navigation stack would need a custom controller and platform-specific kinematic constraints. This is not equivalent to simply swapping the robot base in Nav2.
Control and simulation implications
Rear steering affects both low-level and high-level software. At low level, the controller must synchronize steering position with wheel traction velocity. At high level, planners should respect nonholonomic constraints and steering limits. If these limits are ignored, the robot may command infeasible curvature transitions.
In simulation, rear steering should be represented explicitly in URDF and in the physics model. The joint state stream must include rear steering joint positions. Typical ROS 2 topics include:
/joint_states– steering joint angles and wheel rotation states./odom– estimated body motion./tfand/tf_static– frame transforms according to REP 105./cmd_velor custom controller command topic – body velocity target.
For robot simulation in Gazebo or similar tools, incorrect wheel contact settings can produce unrealistic lateral slip. That problem is more visible on steering-axle platforms than on simple differential-drive robots.
Limitations and trade-offs
A steering rear axle improves mobility, but it adds hardware and software cost. The chassis needs more actuators, steering linkages, encoders, and calibration steps. Odometry is also more sensitive to steering angle bias and linkage backlash.
Main trade-offs include:
- better turning performance versus higher mechanical complexity,
- lower tire scrub versus more demanding calibration,
- more accurate path shaping versus harder controller tuning,
- greater mobility in tight terrain versus more failure points.
For small educational robots, these trade-offs often favor differential drive. For larger research UGVs, especially those carrying heavier sensor payloads, rear steering can be justified if the application requires tighter maneuvering or reduced soil disturbance.
Normative references and standards
In ROS-based robotics work, the most relevant references are REP 103 for standard SI units and coordinate conventions, and REP 105 for mobile robot coordinate frames. URDF and ros2_control documentation define how steering joints and actuator interfaces are represented in ROS 2 systems. For chassis modeling and control, kinematic formulations are commonly aligned with robotics and vehicle dynamics literature rather than with a single ROS-specific standard.
See also
- Ackermann Steering Geometry
- Differential Drive
- URDF
- ROS 2 Control