The Double Pendulum: Golf’s Simplest Useful Model
A full model of the human body has 40+ degrees of freedom. Your spine bends. Your hips rotate. Your knees flex. Your ankles adjust. Even just the arm has many possible movements.
Yet the golf swing has a surprising structure: most of the action happens in two places: shoulder rotation and elbow bend. The torso twists, yes, but it rotates about a fixed vertical axis. The hips move, but mostly to stabilize. The wrist cocks and releases, but its effect on the clubhead is small compared to the arm’s rotation.
So we model the arm as a double pendulum: two rigid links hinged together, hanging from a fixed point (the shoulder). This is a radical simplification. But here is what makes it useful: many of the essential physics of the golf swing show up in this simple model. And because it is simple, we can solve it — find exact equations, compute forces, and understand the dynamics in full.
Once you understand the double pendulum, adding complexity (more joints, torso rotation, etc.) is straightforward.
The Double Pendulum: Physical Setup
The System
Two rigid rods: - Link 1 (upper arm): length \(L_1\), mass \(M_1\), moment of inertia about the shoulder \(I_1\). - Link 2 (forearm + club): length \(L_2\), mass \(M_2\), moment of inertia about the grip \(I_2\).
Two hinges (frictionless, perfect revolutes): - Hinge 1 at the shoulder: the first link rotates about this fixed point. - Hinge 2 at the elbow: the second link rotates about the end of the first link.
Generalized coordinates:
\[ \bm{q} = \begin{pmatrix} \theta_1 \\ \theta_2 \end{pmatrix} \]
where \(\theta_1\) is the angle of link 1 (measured from some reference, say vertical), and \(\theta_2\) is the relative angle of link 2 with respect to link 1.
There are two conventions for defining \(\theta_2\): 1. Absolute: \(\theta_2\) is the angle of link 2 measured from a fixed reference (say, vertical). This is intuitive but makes the kinetic energy more complicated. 2. Relative: \(\theta_2\) is the angle between link 2 and link 1. This makes the kinetic energy simpler because each link rotates about its own hinge.
In this book, we use relative angles. So \(\theta_2 = 0\) means the two links are aligned (arm fully extended). \(\theta_2 = -90°\) means the forearm is bent perpendicular to the upper arm.
Numerical Parameters (Realistic Golf)
Let’s use numbers that match an actual golfer:
| Parameter | Link 1 (Upper Arm) | Link 2 (Forearm + Club) |
|---|---|---|
| Length \(L\) | 0.35 m | 1.0 m |
| Mass \(M\) | 2.5 kg | 1.5 kg |
| CM distance from hinge | 0.175 m | 0.5 m |
| Moment of inertia \(I\) | 0.077 kg m\(^2\) | 0.4 kg m\(^2\) |
These are approximate values for a typical golfer and serve as the canonical two-link baseline for this chapter. Golfers with larger bodies or longer clubs will have different values, but the physics is the same.
The Kinetic Energy
To derive the equations of motion, we start with energy. Every moving system has kinetic energy.
Kinetic Energy of Link 1
Link 1 rotates about the shoulder. Its kinetic energy is:
\[ T_1 = \frac{1}{2} I_1 \dot{\theta}_1^2 \]
where \(I_1 = M_1 L_{1,\text{cm}}^2\) is the moment of inertia of link 1 about the shoulder hinge. (For simplicity, we treat each link as a point mass at its center of mass.)
Kinetic Energy of Link 2
Link 2 rotates about the elbow, which is itself moving. This is trickier.
The elbow is at position:
\[ \bm{r}_{\text{elbow}} = L_1 (\sin \theta_1, -\cos \theta_1) \]
Its velocity is:
\[ \dot{\bm{r}}_{\text{elbow}} = L_1 \dot{\theta}_1 (\cos \theta_1, \sin \theta_1) \]
The center of mass of link 2 is at distance \(L_{2,\text{cm}}\) from the elbow, at angle \(\theta_1 + \theta_2\):
\[ \bm{r}_{\text{cm},2} = L_1 (\sin \theta_1, -\cos \theta_1) + L_{2,\text{cm}} (\sin(\theta_1 + \theta_2), -\cos(\theta_1 + \theta_2)) \]
Its velocity is:
\[ \begin{aligned} \dot{\bm{r}}_{\text{cm},2} &= L_1 \dot{\theta}_1 (\cos \theta_1, \sin \theta_1) \\ &\quad + L_{2,\text{cm}} (\dot{\theta}_1 + \dot{\theta}_2) (\cos(\theta_1 + \theta_2), \sin(\theta_1 + \theta_2)) \end{aligned} \]
The kinetic energy of link 2 is:
\[ T_2 = \frac{1}{2} M_2 |\dot{\bm{r}}_{\text{cm},2}|^2 + \frac{1}{2} I_2 (\dot{\theta}_1 + \dot{\theta}_2)^2 \]
Expanding \(|\dot{\bm{r}}_{\text{cm},2}|^2\):
\[ \begin{aligned} |\dot{\bm{r}}_{\text{cm},2}|^2 &= \big[ L_1 \dot{\theta}_1 \cos \theta_1 + L_{2,\text{cm}} (\dot{\theta}_1 + \dot{\theta}_2) \cos(\theta_1 + \theta_2) \big]^2 \\ &\quad + \big[ L_1 \dot{\theta}_1 \sin \theta_1 + L_{2,\text{cm}} (\dot{\theta}_1 + \dot{\theta}_2) \sin(\theta_1 + \theta_2) \big]^2 \end{aligned} \]
After expanding and simplifying (using \(\cos^2 + \sin^2 = 1\) and \(\cos \alpha \cos \beta + \sin \alpha \sin \beta = \cos(\alpha - \beta)\)):
\[ \begin{aligned} |\dot{\bm{r}}_{\text{cm},2}|^2 &= L_1^2 \dot{\theta}_1^2 + L_{2,\text{cm}}^2 (\dot{\theta}_1 + \dot{\theta}_2)^2 \\ &\quad + 2 L_1 L_{2,\text{cm}} \dot{\theta}_1 (\dot{\theta}_1 + \dot{\theta}_2) \cos \theta_2 \end{aligned} \]
So:
\[ \begin{aligned} T_2 &= \frac{1}{2} M_2 \big[ L_1^2 \dot{\theta}_1^2 + L_{2,\text{cm}}^2 (\dot{\theta}_1 + \dot{\theta}_2)^2 + 2 L_1 L_{2,\text{cm}} \dot{\theta}_1 (\dot{\theta}_1 + \dot{\theta}_2) \cos \theta_2 \big] \\ &\quad + \frac{1}{2} I_2 (\dot{\theta}_1 + \dot{\theta}_2)^2 \end{aligned} \]
Total Kinetic Energy
\[ \begin{aligned} T &= T_1 + T_2 \\ &= \frac{1}{2} I_1 \dot{\theta}_1^2 + \frac{1}{2} M_2 \Big[ L_1^2 \dot{\theta}_1^2 + L_{2,\text{cm}}^2 (\dot{\theta}_1 + \dot{\theta}_2)^2 \\ &\quad + 2 L_1 L_{2,\text{cm}} \dot{\theta}_1 (\dot{\theta}_1 + \dot{\theta}_2) \cos \theta_2 \Big] + \frac{1}{2} I_2 (\dot{\theta}_1 + \dot{\theta}_2)^2 \end{aligned} \]
Collecting terms:
\[ \begin{aligned} T &= \frac{1}{2} (I_1 + M_2 L_1^2) \dot{\theta}_1^2 \\ &\quad + \frac{1}{2} (M_2 L_{2,\text{cm}}^2 + I_2) (\dot{\theta}_1 + \dot{\theta}_2)^2 \\ &\quad + M_2 L_1 L_{2,\text{cm}} \dot{\theta}_1 (\dot{\theta}_1 + \dot{\theta}_2) \cos \theta_2 \end{aligned} \tag{1}\]
Kinetic energy is motion’s “fuel.” In a golf swing, you store energy during the backswing (lifting your arms against gravity, tensioning muscles). You release that energy during the downswing. The kinetic energy increases—the club is moving faster.
By the time the club reaches impact, all that stored energy has been converted to the motion of the club.
The Lagrangian formalism (which we’ll use next) exploits the fact that nature minimizes the action, \(\int (T - V) dt\). By tracking \(T\), we can derive the forces without writing down force balances explicitly.
The Potential Energy
Gravity does work on your arm and the club. This is captured by gravitational potential energy.
The potential energy of a mass at height \(h\) is \(V = Mgh\) (relative to some reference).
For link 1, the center of mass is at height:
\[ h_1 = -L_{1,\text{cm}} \cos \theta_1 \]
(Taking the shoulder as the reference, and negative because we measure downward as negative.)
For link 2, the center of mass is at height:
\[ h_2 = -L_1 \cos \theta_1 - L_{2,\text{cm}} \cos(\theta_1 + \theta_2) \]
Total potential energy:
\[ \begin{aligned} V &= M_1 g h_1 + M_2 g h_2 \\ &= -M_1 g L_{1,\text{cm}} \cos \theta_1 - M_2 g \big[ L_1 \cos \theta_1 + L_{2,\text{cm}} \cos(\theta_1 + \theta_2) \big] \end{aligned} \]
As your arm moves, the potential energy changes. At address (arm down), \(V\) is low. At the top of the backswing (arm up), \(V\) is higher. During the downswing, gravity does work and \(V\) decreases, converting to kinetic energy.
This is why golfers don’t need to “create” all the club’s speed. Gravity helps. The change in \(V\) from top to impact is a “gift” from physics.
The Lagrangian and Euler-Lagrange Equations
Now we use the Lagrangian formalism.
The Lagrangian is defined as:
\[\mathcal{L} = T - V\]
It’s the kinetic energy minus the potential energy.
The Euler-Lagrange equations state that the equations of motion are:
\[ \frac{d}{dt} \left( \frac{\partial \mathcal{L}}{\partial \dot{q}_i} \right) - \frac{\partial \mathcal{L}}{\partial q_i} = \tau_i \]
for each coordinate \(q_i\). Here, \(\tau_i\) is the applied torque (from muscles).
This looks abstract. But it’s a recipe: compute two partial derivatives of the Lagrangian, take one time derivative, subtract them, and you get the equation of motion for that coordinate.
Let’s do it for the double pendulum.
The \(\theta_1\) Equation
We need: - \(\frac{\partial \mathcal{L}}{\partial \dot{\theta}_1}\) - \(\frac{\partial \mathcal{L}}{\partial \theta_1}\)
From \(T\) (Equation 1), the terms involving \(\dot{\theta}_1\) are:
\[ \begin{aligned} \frac{\partial T}{\partial \dot{\theta}_1} &= (I_1 + M_2 L_1^2) \dot{\theta}_1 + (M_2 L_{2,\text{cm}}^2 + I_2) (\dot{\theta}_1 + \dot{\theta}_2) \\ &\quad + M_2 L_1 L_{2,\text{cm}} [(\dot{\theta}_1 + \dot{\theta}_2) + \dot{\theta}_1] \cos \theta_2 \\ &= (I_1 + M_2 L_1^2) \dot{\theta}_1 + (M_2 L_{2,\text{cm}}^2 + I_2) (\dot{\theta}_1 + \dot{\theta}_2) \\ &\quad + M_2 L_1 L_{2,\text{cm}} (2 \dot{\theta}_1 + \dot{\theta}_2) \cos \theta_2 \end{aligned} \]
Taking the time derivative:
\[ \begin{aligned} \frac{d}{dt} \left( \frac{\partial T}{\partial \dot{\theta}_1} \right) &= (I_1 + M_2 L_1^2) \ddot{\theta}_1 + (M_2 L_{2,\text{cm}}^2 + I_2) (\ddot{\theta}_1 + \ddot{\theta}_2) \\ &\quad + M_2 L_1 L_{2,\text{cm}} (2 \ddot{\theta}_1 + \ddot{\theta}_2) \cos \theta_2 \\ &\quad - M_2 L_1 L_{2,\text{cm}} (2 \dot{\theta}_1 + \dot{\theta}_2) \dot{\theta}_2 \sin \theta_2 \end{aligned} \]
Now the partial of \(T\) with respect to \(\theta_1\):
\[ \frac{\partial T}{\partial \theta_1} = 0 \]
(since \(T\) doesn’t depend on \(\theta_1\) explicitly, only on \(\dot{\theta}_1\)).
The partial of \(V\) with respect to \(\theta_1\):
\[ \begin{aligned} \frac{\partial V}{\partial \theta_1} &= M_1 g L_{1,\text{cm}} \sin \theta_1 + M_2 g L_1 \sin \theta_1 + M_2 g L_{2,\text{cm}} \sin(\theta_1 + \theta_2) \\ &= (M_1 L_{1,\text{cm}} + M_2 L_1) g \sin \theta_1 + M_2 g L_{2,\text{cm}} \sin(\theta_1 + \theta_2) \end{aligned} \]
So:
\[ \begin{aligned} -\frac{\partial \mathcal{L}}{\partial \theta_1} &= -\frac{\partial (T - V)}{\partial \theta_1} = \frac{\partial V}{\partial \theta_1} \end{aligned} \]
The Euler-Lagrange equation becomes:
\[ \begin{aligned} &(I_1 + M_2 L_1^2) \ddot{\theta}_1 + (M_2 L_{2,\text{cm}}^2 + I_2) (\ddot{\theta}_1 + \ddot{\theta}_2) \\ &+ M_2 L_1 L_{2,\text{cm}} (2 \ddot{\theta}_1 + \ddot{\theta}_2) \cos \theta_2 \\ &- M_2 L_1 L_{2,\text{cm}} (2 \dot{\theta}_1 + \dot{\theta}_2) \dot{\theta}_2 \sin \theta_2 \\ &= (M_1 L_{1,\text{cm}} + M_2 L_1) g \sin \theta_1 + M_2 g L_{2,\text{cm}} \sin(\theta_1 + \theta_2) + \tau_1 \end{aligned} \tag{2}\]
Similarly, for \(\theta_2\):
\[ \begin{aligned} &(M_2 L_{2,\text{cm}}^2 + I_2) (\ddot{\theta}_1 + \ddot{\theta}_2) + M_2 L_1 L_{2,\text{cm}} \ddot{\theta}_1 \cos \theta_2 \\ &+ M_2 L_1 L_{2,\text{cm}} \dot{\theta}_1^2 \sin \theta_2 \\ &= M_2 g L_{2,\text{cm}} \sin(\theta_1 + \theta_2) + \tau_2 \end{aligned} \tag{3}\]
Equation 2 and Equation 3 are the equations of motion for the double pendulum. They’re complicated, with many terms. Let’s parse them:
- Left side: terms with \(\ddot{\theta}\) are the inertial forces. They say: “to accelerate the arm, you need a torque proportional to the moment of inertia and the acceleration.”
- Gravity terms on the right: e.g., \((M_1 L_{1,\text{cm}} + M_2 L_1) g \sin \theta_1\). These are the gravitational torques. When \(\theta_1 = 0°\) (arm down), \(\sin \theta_1 = 0\), so gravity exerts no torque—the arm is in equilibrium. When \(\theta_1 = 90°\) (arm horizontal), \(\sin \theta_1 = 1\), so gravity exerts maximum torque, trying to pull the arm back down.
- The \(\dot{\theta}_2 \sin \theta_2\) terms: these are Coriolis forces, arising from the coupling between the two joints. When the elbow bends (\(\dot{\theta}_2 \neq 0\)) while the shoulder is rotating (\(\dot{\theta}_1 \neq 0\)), there’s a cross term that affects the dynamics.
- The \(\dot{\theta}_1^2 \sin \theta_2\) term: this is a centrifugal force. When the shoulder rotates fast (\(\dot{\theta}_1\) large), it creates an outward force on the elbow, trying to straighten the arm.
These equations are complicated because the double pendulum is a coupled system. The motion of one joint affects the other.
Rewriting in Matrix Form
The Euler-Lagrange equations are hard to parse. Let’s rewrite them in the standard form:
\[ \bm{M}(\bm{q}) \ddot{\bm{q}} + \bm{C}(\bm{q}, \dot{\bm{q}}) \dot{\bm{q}} + \bm{g}(\bm{q}) = \bm{\tau} \]
where: - \(\bm{M}(\bm{q})\) is the mass matrix (moment of inertia matrix). - \(\bm{C}(\bm{q}, \dot{\bm{q}})\) contains the Coriolis and centrifugal terms. - \(\bm{g}(\bm{q})\) is the gravitational torque. - \(\bm{\tau}\) is the applied (muscular) torque.
The Mass Matrix
Collecting all terms with \(\ddot{\theta}_1\) and \(\ddot{\theta}_2\):
\[ \bm{M}(\bm{q}) = \begin{bmatrix} M_{11}(\theta_2) & M_{12}(\theta_2) \\ M_{21}(\theta_2) & M_{22} \end{bmatrix} \]
where:
\[ \begin{aligned} M_{11}(\theta_2) &= I_1 + M_2 L_1^2 + M_2 L_{2,\text{cm}}^2 + I_2 + 2 M_2 L_1 L_{2,\text{cm}} \cos \theta_2 \\ M_{12}(\theta_2) &= M_{21}(\theta_2) = M_2 L_{2,\text{cm}}^2 + I_2 + M_2 L_1 L_{2,\text{cm}} \cos \theta_2 \\ M_{22} &= M_2 L_{2,\text{cm}}^2 + I_2 \end{aligned} \]
Key observation: \(M_{11}\) depends on \(\theta_2\) (the elbow angle). This means the effective inertia at the shoulder depends on whether the elbow is bent. An extended arm (\(\theta_2 \approx 0\)) has higher inertia than a bent arm (\(\theta_2\) large and negative).
The Coriolis/Centrifugal Matrix
Collecting all terms with \(\dot{\theta}_1 \dot{\theta}_2\) and \(\dot{\theta}^2\):
\[ \bm{C}(\bm{q}, \dot{\bm{q}}) = \begin{bmatrix} 0 & C_{12}(\bm{q}, \dot{\bm{q}}) \\ C_{21}(\bm{q}, \dot{\bm{q}}) & 0 \end{bmatrix} \]
where:
\[ \begin{aligned} C_{12}(\bm{q}, \dot{\bm{q}}) &= -M_2 L_1 L_{2,\text{cm}} (2 \dot{\theta}_1 + \dot{\theta}_2) \sin \theta_2 \\ C_{21}(\bm{q}, \dot{\bm{q}}) &= M_2 L_1 L_{2,\text{cm}} \dot{\theta}_1 \sin \theta_2 \end{aligned} \]
More precisely, the standard form is:
\[ \bm{C}(\bm{q}, \dot{\bm{q}}) \dot{\bm{q}} = \begin{bmatrix} C_{12} \dot{\theta}_2 \\ C_{21} \dot{\theta}_1 \end{bmatrix} \]
But it’s more standard to write the Coriolis matrix such that \(\bm{C} \dot{\bm{q}}\) gives the full velocity-dependent torque. We’ll use the convention where the \((i,j)\) entry couples the velocity of joint \(j\) to joint \(i\).
The Gravity Vector
\[ \bm{g}(\bm{q}) = \begin{bmatrix} (M_1 L_{1,\text{cm}} + M_2 L_1) g \sin \theta_1 + M_2 g L_{2,\text{cm}} \sin(\theta_1 + \theta_2) \\ M_2 g L_{2,\text{cm}} \sin(\theta_1 + \theta_2) \end{bmatrix} \]
Numerical Example: A Realistic Golf Swing
Let’s plug in numbers and see what the forces actually are.
At Impact
Suppose at impact, the arm is nearly extended: - \(\theta_1 = 0°\) (arm pointing toward target) - \(\theta_2 = -5°\) (slight wrist bend) - \(\dot{\theta}_1 = 600°/s = 10.47 \text{ rad/s}\) (shoulder rotation speed) - \(\dot{\theta}_2 = 500°/s = 8.73 \text{ rad/s}\) (elbow/wrist straightening speed)
The mass matrix at this configuration:
\[ \begin{aligned} M_{11} &\approx 0.077 + 1.5 \cdot 0.35^2 + 1.5 \cdot 0.5^2 + 0.4 + 2 \cdot 1.5 \cdot 0.35 \cdot 0.5 \cos(-5°) \\ &\approx 0.077 + 0.184 + 0.375 + 0.4 + 0.523 \\ &\approx 1.56 \text{ kg m}^2 \end{aligned} \]
(This is the effective moment of inertia of the entire arm+club system about the shoulder.)
The gravitational torques:
\[ \begin{aligned} g_1 &= (2.5 \cdot 0.175 + 1.5 \cdot 0.35) \cdot 9.81 \cdot \sin(0°) + 1.5 \cdot 9.81 \cdot 0.5 \cdot \sin(-5°) \\ &\approx 0 - 0.64 \text{ N m} \\ &\approx -0.64 \text{ N m} \end{aligned} \]
(Negative means gravity is pulling the arm downward/backward.)
The Coriolis term at the elbow:
\[ \begin{aligned} C_{21} \dot{\theta}_1 &= M_2 L_1 L_{2,\text{cm}} \dot{\theta}_1^2 \sin \theta_2 \\ &= 1.5 \cdot 0.35 \cdot 0.5 \cdot 10.47^2 \cdot \sin(-5°) \\ &\approx 1.5 \cdot 0.35 \cdot 0.5 \cdot 109.7 \cdot (-0.087) \\ &\approx -2.5 \text{ N m} \end{aligned} \]
At impact: - The total mass/inertia of the arm+club is about 1.56 kg m\(^2\). - Gravity is pulling with a torque of 0.64 N m. - The centrifugal term (from the high shoulder rotation) is creating a torque of 2.5 N m trying to straighten the elbow.
For comparison, a human can generate roughly 30–50 N m of torque at the shoulder (Nesbit 2005; Gatt et al. 1998) (illustrative; values vary with body size, training, and joint angle). So gravity and centrifugal effects are much smaller than the muscular torque budget in this illustrative example. The muscles do the rest.
But notice: if we had generated all of this torque during the downswing, the club would barely accelerate. Instead, we generate a moderate muscular torque and let gravity and momentum contribute. That’s efficiency.
The Interaction Torque: How the Arm Pulls the Club
One of the most important terms is the interaction torque between the two links.
In the equations of motion, we have the term \(M_2 L_1 L_{2,\text{cm}} \cos \theta_2\) appearing in \(M_{11}\). This couples the two joints. It means: the acceleration of the shoulder depends on the angle of the elbow.
The torque that one joint exerts on the other through the kinetic energy coupling. In the double pendulum, the term \(M_2 L_1 L_{2,\text{cm}} (\ddot{\theta}_1 + \ddot{\theta}_2) \cos \theta_2\) in the \(\theta_1\) equation is the torque that the elbow exerts on the shoulder through their coupling.
Why does this matter? Because when you accelerate the elbow (throw the club down), it pulls on the upper arm. This is why you feel the club being “heavy” during the downswing. It’s not that the club itself is heavy—it’s that accelerating it creates a reaction torque that pulls on your shoulder.
Example: The Whip Effect
Imagine cracking a whip. The handle accelerates slowly, the middle segment accelerates faster, and the tip accelerates fastest. Each segment pulls the segment behind it.
In the golf swing, the same thing happens. The shoulder rotates at some speed. The elbow catches up and rotates faster. In doing so, it pulls on the shoulder through the interaction torque. This is why amateur golfers often feel the club dragging—they’re not accelerating the elbow fast enough to create that interaction torque.
Elite golfers feel the club pulling—they’ve trained to accelerate the elbow hard, which pulls through the shoulder and wrist, creating a whip effect.
Chaotic Dynamics: Why Small Differences Matter
Here’s a fascinating property of the double pendulum: it can exhibit chaotic behavior.
Motion is chaotic if small changes in initial conditions lead to drastically different outcomes, despite the system being completely deterministic.
For the golf swing, this has profound implications. Two golfers with nearly identical swings at the top (same angles and velocities) might produce wildly different impact conditions if the equations of motion are chaotic.
Actually, in the absence of control (\(\bm{\tau} = 0\)), the double pendulum exhibits chaos only in certain regimes. But once you add muscular control (\(\bm{\tau} \neq 0\)), the picture becomes complex. Small errors in muscle activation can compound.
This is why consistency is hard. It’s not just about executing a “swing shape”—it’s about executing precise forces at precise times.
Here’s a simple example. Suppose your shoulder is rotating at 10.47 rad/s and you want to accelerate the club by releasing the wrist (\(\ddot{\theta}_2 > 0\)).
If you release perfectly, the club accelerates and hits the sweet spot.
If you release just slightly late (by 0.01 seconds), the coupling terms in the equations mean the shoulder and elbow positions have evolved differently. The wrist is now at a slightly different angle and position. The release now happens at a different configuration. The club’s trajectory is different. Impact happens at a different location on the clubface. The ball flies to a different target.
Over the course of a full swing with multiple muscles controlling multiple DOF, these small errors can accumulate. This is why practice is essential: you’re training your nervous system to execute precise control signals, overcoming the inherent chaos of the dynamics.
From Force Equations to Physical Insight
The manipulator equation \(\bm{M} \ddot{\bm{q}} + \bm{C} \dot{\bm{q}} + \bm{g} = \bm{\tau}\) is not just mathematics—it’s a statement about physics.
Left side = what the system wants to do given its current state (gravity pulling, momentum carrying it forward, inertia resisting acceleration).
Right side = what you’re actively commanding with muscles.
The difference between amateurs and experts isn’t just that experts generate larger forces. It’s that experts understand the left side and use small commands on the right to steer the passive dynamics toward their goal.
- The double pendulum captures the essential physics of the golf swing: two linked segments with gravity, inertia, and velocity-dependent forces.
- The kinetic energy depends on the speeds and configuration. An extended arm has higher inertia than a bent one. This is why golfers sometimes lengthen the swing (to increase the moment arm and feel the club’s inertia).
- The mass matrix \(M(\theta_2)\) couples the joints. The effective inertia at the shoulder depends on the elbow angle.
- Gravity exerts torques that vary with configuration. At address (arm down), \(\sin \theta_1 = 0\), so gravity is balanced. At the top (arm up), gravity pulls strongly.
- Coriolis and centrifugal forces are huge in golf. At 100 mph club speed, the centrifugal forces are on the order of \(100 g\).
- The interaction torque couples the two joints. Accelerating the elbow creates a reaction torque at the shoulder. This is the “whip effect.”
- Small differences in control can lead to large differences in outcome. The dynamics are sensitive to initial conditions. Consistency requires precise muscular control.
Chapter Exercises
- Moment of Inertia. A rod of mass \(M\) and length \(L\) has moment of inertia \(I = \frac{1}{12} M L^2\) about its center and \(I = \frac{1}{3} M L^2\) about one end. For the forearm+club (mass 1.5 kg, length 1.0 m), compute the moment of inertia about the elbow.
- Potential Energy Change. At address, both arms are down (\(\theta_1 = 0°, \theta_2 = 0°\)). At the top of the backswing, \(\theta_1 = 150°, \theta_2 = -70°\). Using the parameters in Section 1.1.2, compute the change in potential energy. How much gravitational work is available during the downswing?
- Centrifugal Force. At impact, \(\dot{\theta}_1 = 10.47 \text{ rad/s}\). The centrifugal acceleration at the clubhead (distance \(L_1 + L_2 = 1.35\) m from shoulder) is \(a_c = \dot{\theta}_1^2 \cdot r\). Compute this acceleration in units of \(g\).
- Impact Accelerations. A golf club at impact often experiences accelerations of \(170g\) or more. Using the double pendulum model, estimate what shoulder and elbow accelerations (\(\ddot{\theta}_1, \ddot{\theta}_2\)) are needed to produce this. (Hint: the clubhead acceleration is \(a = L_1 \ddot{\theta}_1 + \text{other terms}\).)
- Gravity Torque. At the top of the backswing (\(\theta_1 = 150°, \theta_2 = -70°\)), compute the gravitational torque at the shoulder. Is gravity trying to pull the arm backward or forward? Why?
- Chaotic Dynamics. Look up the properties of the (unforced) double pendulum. In what regime does it exhibit chaos? Does the golf swing operate in that regime?