⚡ electrical engineering

Equilibrium Analysis and Gyro-Stabilization of Two-Wheel Inverted Pendulum Robot on Inclined Surfaces

This study presents a nonlinear dynamic model and a dual-loop control strategy for a two-wheel inverted pendulum robot equipped with a scissor-pair control moment gyroscope, demonstrating its ability to maintain equilibrium and navigate inclined surfaces by leveraging gyroscopic torque to counteract gravitational forces at a specific tilt angle.

Original authors: ِAhmed Amin, Moustafa Fouz, Ahmed Elsawaf

Published 2026-08-07

📖 9 min read🧠 Deep dive

CC BY 4.0

Original authors: ِAhmed Amin, Moustafa Fouz, Ahmed Elsawaf

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a world where robots don't just roll on flat floors but can gracefully navigate up steep hills, balancing on two wheels like a human on a unicycle. This is the realm of the Two-Wheel Inverted Pendulum Robot (TWIPR). Think of it as a mechanical version of the classic "cart-and-pole" problem: a heavy stick (the robot's body) is balanced on top of a moving cart (the wheels). If the stick leans even a tiny bit, gravity wants to pull it down, so the cart must constantly move back and forth to catch it. This is tricky because the robot is "underactuated," meaning it has fewer controls than it has ways to move; it can't just push a button to stay upright. Usually, these robots rely entirely on their wheels spinning faster or slower to stay balanced. But what happens when the ground isn't flat? On a hill, gravity pulls the robot down the slope, making the balancing act even harder. To solve this, scientists have started looking at gyroscopes—spinning wheels that act like stubborn, heavy tops. Just like a spinning bicycle wheel resists being tilted, a gyroscope inside a robot can generate a powerful twisting force to help it stay upright without the wheels having to do all the heavy lifting.

This paper explores a clever new way to build such a robot for hilly terrain. The researchers, Ahmed Amin, Moustafa Fouz, and Ahmed Elsawaf, designed a robot that uses a special "scissor-pair" of gyroscopes to help it balance. They discovered that on a slope, the robot can't just stand perfectly straight up; it has to lean at a specific, non-zero angle to stay in balance, much like a cyclist leaning into a turn. They created a detailed mathematical model to predict exactly how much the robot needs to lean based on the steepness of the hill and the robot's weight distribution. They also proved that while the gyroscopes are great for quick corrections, they can't hold the robot up forever on a steep hill; eventually, the robot's wheels must provide the steady force to fight gravity. To test their ideas, they built a virtual robot in a computer simulation and found that their math matched the computer's physics perfectly. Their proposed control system, which uses one brain for speed and another for balance, successfully kept the robot upright and moving on a 5-degree slope, proving that this gyro-assisted design is a viable way to tackle inclined surfaces.

The Robot with a Spinning Secret

Meet the robot: a two-wheeled machine that looks like a futuristic scooter but has a secret weapon hidden inside its chest. Instead of just relying on its wheels to stay upright, this robot carries a pair of high-speed spinning flywheels, known as a Scissor-Pair Control Moment Gyroscope (SPCMG). Imagine two spinning tops inside the robot, mounted on a frame that can tilt them back and forth. When the robot starts to wobble, these tops don't just spin; they tilt, creating a powerful "gyroscopic torque" that pushes the robot back into balance. It's like having a magical internal muscle that can twist the robot's body without touching the ground.

The researchers wanted to see if this magic muscle could help the robot handle a real challenge: inclined surfaces. Most robots are tested on flat floors, but the real world is full of ramps and hills. The team asked: Can this gyro-stabilized robot balance on a slope? If so, how does it do it, and what are the limits?

The Tilted Truth: Why Straight Up Doesn't Work

One of the paper's most interesting findings is a bit counterintuitive. You might think that to balance on a hill, a robot should stand perfectly straight up. But the math says otherwise. The researchers found that on a slope, the robot must lean at a specific angle to stay in equilibrium.

Think of it like this: If you are standing on a steep hill, you naturally lean forward to keep from falling backward. The robot does the same. The team derived a formula showing that the angle the robot leans (the equilibrium tilt angle) depends on two things: how steep the hill is and a special number they called the Chassis-to-Robot Static Torque Ratio (represented by the Greek letter $\xi$ ). This ratio is basically a measure of how the robot's weight is distributed compared to its wheel size.

The paper shows that if the robot's design has a low $\xi$ value, it has to lean over a lot to stay balanced on a hill. If the design has a high $\xi$ , it can stand more upright. This relationship is crucial because it tells engineers exactly how to build the robot so it doesn't tip over on a slope. The researchers calculated that for their specific robot design, with a $\xi$ of 0.5, standing on a 5-degree slope requires the robot to lean at an equilibrium angle of about 10 degrees.

The Limits of the Spin: When Gyros Need Help

Here is where the story gets a bit more grounded. While the gyroscopes are amazing for quick, sudden movements (like catching the robot if it starts to fall), the paper explicitly rules out the idea that they can hold the robot up forever on a steep hill.

The gyroscopes work by tilting their spinning wheels. But they can only tilt so far before they hit a "singularity"—a point where they lose their power to push back. It's like trying to push a door open when you are standing right next to the hinge; eventually, you run out of leverage. The paper explains that the gyroscopes cannot provide a constant, steady force to fight gravity indefinitely. Instead, the robot's wheels must take over the job of holding it up against the slope's pull. The gyroscopes are the "emergency brakes" and "steering wheel," while the wheels are the "engine" that keeps the robot from sliding down.

The researchers also calculated the absolute limits of how steep a hill this robot can handle. They found two main barriers:

Geometry: If the hill is too steep, the robot's body would hit the ground before it could lean enough to balance.
Friction: The wheels need enough grip (friction) to push against the hill without slipping.

Combining these, the paper suggests that the maximum slope the robot can handle is determined by the smaller of two values: the robot's design ratio ( $\xi$ ) or the friction of the ground ( $\mu_s$ ). For their robot, this means it can theoretically handle slopes up to a certain limit, but if the ground is slippery or the design is "top-heavy," that limit drops.

The Brain Trust: Two Controllers Working Together

To make all this happen, the team didn't just build the robot; they built a brain for it. They used a "hierarchical" control system, which is like having two managers working together.

Manager 1 (The Locomotion Controller): This manager uses a PID controller (a standard type of control system) to tell the wheels how fast to spin. Its only job is to make sure the robot goes the speed the user wants, whether that's speeding up, slowing down, or stopping.
Manager 2 (The Balance Controller): This manager uses a fancy algorithm called LQR (Linear Quadratic Regulator). Its job is to watch the robot's tilt and the gyroscope's angle. If the robot starts to wobble, this manager tells the gyroscopes to tilt and push the robot back to the correct leaning angle.

The beauty of this system is that they work independently but in sync. The wheels handle the speed, and the gyroscopes handle the balance. The researchers tested this in a computer simulation (using MATLAB/Simulink) and found that it worked beautifully.

The Simulation: A 30-Second Test Run

To prove their ideas, the team ran a simulation where the robot started on a 5-degree slope. They gave it a command: "Go!" The robot accelerated to a speed of 1 m/s.

The Result: The robot reached that speed in about 2.6 seconds without wobbling out of control.
The Stop: Then, they told it to stop. The robot slowed down and came to a halt in about 3.8 seconds.

During these maneuvers, the robot's body tilted back and forth a bit (reaching a peak tilt of about 26.15 degrees when it first started moving), but the gyroscopes quickly corrected it, bringing it back to the steady 10-degree lean required for that slope. The gyroscopes themselves only tilted about 32.5 degrees at their most extreme, which is well within their safe limits (they can go up to 90 degrees before losing power).

The paper also checked the forces involved. To keep the robot moving at a steady 1 m/s up the hill, the wheels needed to apply a constant torque of about 0.386 N·m just to fight gravity. When the robot was speeding up, the wheels had to push harder, reaching about 1.5 N·m. When stopping, they had to brake, reaching about -1.1 N·m. The gyroscopes, meanwhile, only had to work hard during the quick changes in speed, providing a maximum torque of about 1.45 N·m to keep the robot from falling over.

The Verdict

The paper concludes that this gyro-stabilized design is a solid, mathematically sound way to build robots that can handle hills. The computer simulations showed that the robot could balance, move, and stop on a slope without falling over, and the math they wrote down perfectly predicted what the computer simulation did.

However, the authors are careful to note that this is currently a simulation. They haven't built a physical robot to test this on a real hill yet. They suggest that the next step is to build a prototype and see if it works in the real world, where things like slippery leaves or bumpy ground might make things harder than in the computer. But for now, the math looks good, the gyroscopes seem ready, and the path to hill-climbing robots is looking a lot clearer.

Technical Summary: Equilibrium Analysis and Gyro-Stabilization of Two-Wheel Inverted Pendulum Robot on Inclined Surfaces

Problem Statement
Two-wheel inverted pendulum robots (TWIPRs) are inherently unstable, underactuated systems where balance is traditionally achieved solely through wheel torque. While effective on flat terrain, conventional TWIPRs face significant challenges on inclined surfaces. The coupling between translational motion and body balance complicates control, and existing studies on slope operation are largely limited to wheel-driven architectures or rely on linearized models that fail to capture the dynamics of systems with internal momentum-exchange mechanisms. Furthermore, the integration of Control Moment Gyroscopes (CMGs) for stabilization introduces new complexities: CMGs cannot sustain constant torque in a fixed direction indefinitely due to gimbal singularities, and the relationship between terrain inclination, structural parameters, and the resulting equilibrium posture in gyro-stabilized systems remains unexplored.

Methodology
The authors propose a comprehensive analytical and control framework for a TWIPR equipped with a Scissor-Pair Control Moment Gyroscope (SPCMG) operating on inclined surfaces.

Dynamic Modeling: A complete nonlinear dynamic model is derived using the Euler–Lagrange formulation. The model incorporates the coupled dynamics of wheel-driven motion, chassis tilt, and internal gyroscopic actuation, explicitly accounting for slope-induced gravitational effects. The system is defined by generalized coordinates representing longitudinal displacement ( $z$ ), chassis tilt angle ( $\phi$ ), and gimbal precession angle ( $\theta$ ).
Model Validation: To ensure accuracy, a high-fidelity multibody model is developed in MATLAB/Simulink/Simscape Multibody. This physics-based representation serves as an independent verification tool against the derived mathematical equations.
Equilibrium Analysis: The study derives an analytical expression for the slope-dependent equilibrium configuration. It introduces a dimensionless metric, the Chassis-to-Robot Static Torque Ratio ( $\xi$ ), defined as $\xi = \frac{m_c L}{m_R r}$ . This parameter links mass distribution and geometry to the equilibrium tilt angle. The analysis determines the maximum admissible terrain inclination by combining geometric constraints (chassis collision limits) and wheel-surface traction constraints (friction limits).
Control Architecture: A decoupled hierarchical control strategy is implemented:
- Locomotion Subsystem: A Proportional–Integral–Derivative (PID) controller regulates longitudinal velocity by modulating wheel torque.
- Balance Subsystem: A Linear Quadratic Regulator (LQR) controls the gimbal angular velocity to stabilize the chassis around the slope-dependent equilibrium angle. The LQR design includes an integral state for the gimbal angle to ensure it returns to zero (neutral position) after transient disturbances, preserving actuation authority.

Key Contributions

Nonlinear Dynamic Model: Derivation of a full nonlinear model for a gyro-stabilized TWIPR on inclined surfaces, capturing the interaction between gyroscopic coupling and slope-induced gravity.
Equilibrium Analysis: Establishment of an analytical relationship between surface inclination and equilibrium chassis tilt. The introduction of the Chassis-to-Robot Static Torque Ratio ( $\xi$ ) provides a generalized design metric for evaluating slope-handling capabilities and predicting the maximum feasible operating inclination.
Unified Inclination Limit: A derived criterion ( $\alpha_{max} = \tan^{-1}(\min(\xi, \mu_s))$ ) that defines the maximum stable operating angle based on both mechanical design ( $\xi$ ) and terrain friction ( $\mu_s$ ).
Control Framework: Development of a PID-LQR control scheme that successfully manages the dual objectives of velocity tracking and balance stabilization around a non-zero equilibrium tilt, ensuring the gimbal returns to a neutral state to avoid singularities.

Results
The proposed framework was validated through simulations comparing the nonlinear mathematical model and the multibody Simulink model on a $5^\circ$ incline.

Model Accuracy: The simulation results showed strong agreement between the analytical model and the multibody simulation across all dynamic variables, confirming the validity of the derived equations.
Performance: The robot successfully tracked a step speed command (0 to 1 m/s and back to 0) with a settling time of approximately 2.6 seconds and no overshoot during acceleration.
Stabilization: The chassis tilt angle converged to the analytically predicted equilibrium value of approximately $10^\circ$ . During transient maneuvers, the gimbal angle deviated by a maximum of $32.5^\circ$ , remaining well within the safe operating range and far from the $90^\circ$ singularity.
Actuation: The control strategy demonstrated that while wheel torque is required to counteract gravitational loading on the slope, the gyroscopic actuation is primarily active during transients. Once equilibrium is reached, the gimbal torque diminishes to near zero, confirming that the SPCMG does not need to sustain constant torque for steady-state balance.

Significance
The paper claims to provide a validated and practically applicable framework for the analysis and control of gyro-stabilized TWIPRs on inclined terrain. By characterizing the equilibrium behavior through the dimensionless ratio $\xi$ , the work offers a systematic method for designing robots capable of operating on slopes without relying solely on aggressive wheel motion. The study demonstrates that integrating SPCMGs allows for robust balance stabilization while maintaining the ability to return the gimbal to a neutral position, thereby avoiding singularities and preserving control authority. The authors conclude that this approach effectively addresses the limitations of existing linearized models and wheel-only stabilization strategies, offering a foundation for future experimental validation and extension to adaptive control or 3D motion scenarios.