Traditional vehicle steering relies on redirecting ambient fluid flow or changing contact geometry. A ship uses a rudder, an airplane uses ailerons and elevators, and a car turns its front wheels. While effective, these classical mechanisms share a fundamental limitation: control authority is tightly coupled to the vehicle's forward velocity or surface traction. If a boat stops moving, its rudder becomes useless; if a car encounters black ice, its steering wheel loses control.
A rapidly growing body of research circumvents these limitations by shifting the responsibility of steering from passive surfaces to independent and differential actuator control. By treating every thruster, wheel, or rotor as an autonomous source of both force and torque, modern robotic systems can decouple their orientation from their translational path, navigate highly confined environments, and maintain control authority even when environmental traction completely fails.
The Power of the Moment Arm: Differential Actuation
The simplest way to achieve independent steering without mechanical linkages is through differential actuation. When two or more independent propulsion units are separated by a physical distance, they create a moment arm relative to the vehicle's center of mass.
Thrust (F1) ↑ ↑ Thrust (F2)
| ______ |
`-[______]-'
|
Center of Mass
If both actuators produce equal force, the vehicle moves straight ahead. However, modulating their relative outputs generates a net torque directly:
$$\tau = (F_1 - F_2) \cdot \frac{d}{2}$$
where $d$ is the distance between the actuators. As outlined by Madhukar BN et al., this differential thrust is what allows standard quadrotors to precisely control pitch, roll, and yaw purely by varying motor speeds.
This exact principle is being leveraged to eliminate heavy, complex mechanical steering linkages in multi-modal and surface vehicles. For instance, Wang demonstrates a catamaran garbage-cleaning robot that completely discards traditional rudders, relying instead on two independent rear-mounted propellers to execute tight turns in narrow urban waterways.
Remarkably, this mechanism transfers seamlessly across physical domains. Zheng et al. developed a hybrid "unicycle-drone" that rolls on a single large wheel for high-endurance ground locomotion. Instead of using a steering servo, it utilizes differential thrust from its aerial rotors to generate yaw torque on the ground. Similarly, Zheng et al. developed a 37.4g flapping-wing robot that crawls on passive legs; by varying the oscillation frequencies of its independent wing modules on opposite sides, it achieves tank-like differential steering on land using the exact same actuators it uses for flight.
On the ground, Individual Wheel Drive (IWD) systems apply this concept to tire forces. Traditional cars distribute power through a central differential, but Zhou et al. demonstrate that controlling each wheel's speed and torque independently provides unprecedented control authority. By generating precise yaw moments through differential wheel speeds, their system can initiate and stabilize aggressive, high-sideslip drifting maneuvers at the absolute handling limits of the vehicle.
Decoupling Attitude from Translation: Thrust Vectoring
While differential thrust alters a vehicle's orientation by generating torque, it still requires the vehicle to rotate its entire body to redirect its primary thrust vector. If a standard quadrotor wants to move sideways, it must first roll.
To achieve true multi-degree-of-freedom maneuverability, researchers are decoupling thrust vectoring from the vehicle's body attitude by allowing the propulsion units themselves to tilt.
Tiltable Rotor Tiltable Rotor
\ /
\ _________________ /
`-[_________________]-'
In the underwater domain, this capability is critical. The authors of a recent CFD study show that a three-propeller architecture—combining a rear thruster for forward drive with two front-side thrusters—enables agile differential steering and stabilization in confined agricultural ponds. Taking this further, Liu et al. introduced an aerial-aquatic quadrotor where each propulsion unit can independently tilt around its mounting arm. This allows the vehicle to execute agile yaw and horizontal movements underwater without tilting its main body, drastically improving response speed and control flexibility in a dense fluid medium.
This independent tilting mechanism provides two profound advantages:
- Force Redirection: González et al. introduce a vectorial propulsion system for an Autonomous Underwater Vehicle (AUV) that adjusts the angle of lateral thrusters, granting full 3D maneuverability across horizontal and vertical planes for precise depth and attitude control.
- High-Force Interaction: In aerial robotics, Hui et al. utilize tiltable rotors to vector thrust in non-vertical directions. This allows their platform, AEROBULL, to exert massive horizontal forces against walls and structures for industrial inspection without losing hovering stability.
Overcoming Environmental Friction Limits
One of the most compelling reasons to adopt independent actuator control is its ability to bypass the physical limitations of the operating environment.
Traditional wheeled robots rely entirely on surface friction to generate centripetal force during a turn. On slippery or low-traction surfaces like ice, mud, or wet tiles, the wheels slip, and control is lost. Lin et al. solve this with "Skater," a bi-copter robot equipped with passive wheels. By using tilted aerial rotors to generate aerodynamic vectored thrust, the robot produces active lateral forces directly in the air. Because the turning force is decoupled from ground friction, the vehicle maintains stable, high-speed trajectories on surfaces where traditional wheeled steering would completely fail.
Similarly, in highly restricted or hazardous environments, independent propulsion architectures enable zero-radius turns that mechanical steering linkages cannot physically execute. Pairet et al. implement independent, steerable propulsion units on a surface vehicle to achieve precise station-keeping and zero-radius turns in crowded harbors. For even tighter spaces, Nancekievill et al. designed a miniature cylindrical ROV for navigating flooded nuclear containment vessels; by integrating multiple independent water pumps directly into the vehicle's end-caps, they achieved multi-degree-of-freedom movement and a turning radius equal to half the vehicle's length.
The Mathematics of Decoupled Control
From a control theory perspective, managing these independent forces requires moving away from simplified linear models. When position and orientation are tightly coupled, traditional controllers suffer from mathematical singularities and "unwinding" issues during aggressive maneuvers.
To fully exploit independent actuation, Lee et al. developed a robust nonlinear tracking controller designed directly on the Special Euclidean group $SE(3)$. By utilizing a geometric projection method, their controller can isolate and specify a desired heading (yaw) without interfering with primary 3D translational path-following objectives.
Underactuated (Coupled) Fully Decoupled (e.g., ALICE)
Must tilt/turn to sway Direct sway via lateral thrusters
▲ ▲
| |
◄───┼───► ◄───┼───►
| ▲ | ▲
▼ | ▼ |
Lateral thrusters provide
independent lateral force
When a system is physically underactuated, control algorithms must cleverly exploit the vehicle's dynamics to transition between states—such as the pivoting control strategy proposed by Ma et al., which uses thrust vectoring to rotate a tailsitter aircraft from a horizontal belly-sitting position into a vertical hover.
Alternatively, engineers can add dedicated actuators to explicitly decouple the system's equations of motion. Gutnik et al. demonstrated this by modifying a standard torpedo-shaped AUV, which traditionally could only control surge, heave, and limited yaw. By adding independent lateral thrusters, they provided direct control over sway and decoupled yaw entirely from forward velocity. This modification transforms an underactuated vehicle into a fully decoupled platform capable of the intricate, low-altitude lateral movements required for close-range seabed imaging.
Key Foundational Work
A visual servoing system for an aquatic swimming robot (2005) — contributed 1 concept elements to this direction.
Nonlinear Robust Tracking Control of a Quadrotor UAV on SE(3) (2011) — contributed 1 concept elements to this direction.