A robot, in the broadest sense introduced in Chapter 1, is a physical system that executes purposeful motion using internal mechanisms and energy. Long before electronics, automata achieved purpose through mechanical computation: geometry and inertia encoded sequences, and stored energy drove motion. Understanding these devices clarifies three foundational robotics ideas:

• Actuation: converting stored energy into motion.
• Programming: pre-specifying motion by mechanism geometry.
• Regularization: stabilizing periodic motion by timekeeping mechanisms.

flowchart TD
  A["Human desire to imitate life"] --> B["Mechanisms to create motion"]
  B --> C["Energy sources: weights, springs, water, steam"]
  C --> D["Motion encoding: gears, cams, pinned drums"]
  D --> E["Sequenced behavior: automata theaters, clocks"]
  E --> F["Conceptual bridge to robots"]
        

Early automata appeared independently across civilizations, typically linked to temple engineering, entertainment, or timekeeping:

1. Ancient Egypt & Greece (3rd–1st century BCE): water clocks and animated statues; Greek engineers described self-opening doors and moving figures using pneumatics and hydraulics.
2. China (Han to Song dynasties): mechanical birds, humanoid figurines, and clockwork-driven astronomical devices.
3. Islamic Golden Age (9th–13th century): precise water-powered clocks and programmable musical automata (pinned drums).
4. European Renaissance (15th–17th century): spring-driven humanoids and complex clockwork; systematic design of cams and gear trains.

The unifying mechanical theme is predefined motion under open-loop drive, often regulated by escapements (Section 4).

Many automata used gear trains to scale torque and speed. For two meshed gears with tooth counts \( N_1 \) and \( N_2 \), angular velocities satisfy

\[ \omega_2 = -\frac{N_1}{N_2}\,\omega_1, \qquad \tau_2 = -\frac{N_2}{N_1}\,\tau_1. \]

The minus sign indicates opposite rotation. For a gear train with stages \( (1\to2\to3\to\cdots\to m) \),

\[ \frac{\omega_m}{\omega_1} = (-1)^{m-1}\prod_{k=1}^{m-1} \frac{N_k}{N_{k+1}}. \]

Proposition (Rational speed ratios). Any gear train with integer tooth counts yields a rational speed ratio.

Proof. Each stage ratio \(N_k/N_{k+1}\) is a ratio of integers, hence rational. The finite product of rationals is rational. Therefore \( \omega_m/\omega_1 \in \mathbb{Q} \). This matters historically because clockmakers could realize precise fractional ratios for periodic motion without continuous tuning.

Early robotic ancestry: modern reducers and harmonic drives are direct descendants of these ratio-setting ideas.

A cam converts a rotating input into a designed displacement profile. Let cam angle be \( \theta(t) \) with constant speed \( \dot{\theta}=\omega \). A follower displacement profile is a periodic function \( x(\theta) \), so

\[ x(t) = x(\theta(t)), \qquad \theta(t)=\omega t + \theta_0. \]

Most automata used profiles built from simple segments. A common university-level abstraction is a piecewise polynomial cam:

\[ x(\theta)= \begin{cases} a_0 + a_1\theta + a_2\theta^2, & 0 \le \theta < \theta_r \\ x(\theta_r), & \theta_r \le \theta < \theta_h \\ b_0 + b_1\theta + b_2\theta^2, & \theta_h \le \theta < 2\pi \end{cases} \]

where \( \theta_r \) is a rise interval and \( \theta_h \) a hold interval. Coefficients are chosen to enforce smoothness:

\[ \text{continuity: } x(\theta_r^-)=x(\theta_r^+), \quad \text{and often } x'(\theta_r^-)=x'(\theta_r^+). \]

flowchart TD
  R["Rotate cam at speed omega"] --> P["Cam profile x(theta)"]
  P --> F["Follower displacement x(t)"]
  F --> L["Linkages move figure"]
  L --> S["Sequenced behavior repeats every 2pi"]
        

Historically, a cam’s geometry stored the “algorithm.” This is a direct mechanical analogue of later motion planning: a desired path encoded before execution.

Clocks and automata theaters required stable timing. Escapements periodically release energy from a weight or spring to maintain oscillation. A minimal physics model treats the pendulum or balance as a damped oscillator driven by impulses:

\[ I\ddot{\phi}(t) + c\dot{\phi}(t) + k\phi(t) = \sum_{n=0}^{\infty} J\,\delta(t-nT), \]

where \( \phi \) is angular displacement, \( I \) inertia, \( c \) damping, \( k \) restoring stiffness (gravity or spring), \( J \) impulse magnitude, and \( T \) the escapement period.

Stability intuition. Without impulses \(J=0\), energy decays exponentially due to damping. The impulse train replaces the lost energy, creating a stable limit cycle. The average injected power per cycle is \( P_{in} = J\omega_{osc}/(2\pi) \), which balances average dissipated power \( P_{diss} \propto c \langle \dot{\phi}^2\rangle \).

This is a proto-control concept: a mechanism that enforces regularity in time, enabling synchronized motion sequences.

Even though early automata are historical, simple simulations help you recognize their mathematical structure.

6.1 Python — Simulating a Piecewise Cam Profile

import numpy as np
import matplotlib.pyplot as plt

# cam parameters
theta_r = np.pi/2        # rise ends at 90 degrees
theta_h = np.pi          # hold ends at 180 degrees
omega = 2.0              # rad/s

def cam_profile(theta):
    theta = np.mod(theta, 2*np.pi)
    if theta < theta_r:
        # quadratic rise: x = a2 * theta^2
        a2 = 1.0 / theta_r**2
        return a2 * theta**2
    elif theta < theta_h:
        return 1.0
    else:
        # quadratic return to 0 at 2pi
        b2 = 1.0 / (2*np.pi - theta_h)**2
        return b2 * (2*np.pi - theta)**2

thetas = np.linspace(0, 2*np.pi, 500)
xs = np.array([cam_profile(th) for th in thetas])
t = thetas/omega

plt.figure()
plt.plot(t, xs)
plt.xlabel("time (s)")
plt.ylabel("follower displacement x(t)")
plt.title("Cam-driven periodic motion")
plt.grid(True)
plt.show()

6.2 C++ — Gear Ratio Calculator (Tooth-Count Programming)

#include 
#include 

int main() {
    std::vector N = {20, 60, 15, 45}; // tooth counts for gears 1..4
    double ratio = 1.0;
    int sign = 1;

    for (size_t k = 0; k + 1 < N.size(); ++k) {
        ratio *= static_cast(N[k]) / N[k+1];
        sign *= -1;
    }

    std::cout << "Speed ratio omega_m/omega_1 = "
              << sign * ratio << std::endl;
    return 0;
}

6.3 Java — Discrete Escapement Impulse Oscillator

public class EscapementOscillator {
    public static void main(String[] args) {
        double I = 1.0, c = 0.08, k = 4.0;
        double J = 0.15, T = 1.0;  // impulse every T seconds
        double dt = 0.001, tf = 10.0;

        double phi = 0.2, dphi = 0.0;
        for (int step = 0; step * dt < tf; step++) {
            double t = step * dt;

            // impulse at multiples of T
            if (Math.abs(t - Math.round(t / T) * T) < dt / 2.0) {
                dphi += J / I;
            }

            double ddphi = -(c/I)*dphi - (k/I)*phi;
            dphi += ddphi * dt;
            phi += dphi * dt;

            if (step % 100 == 0) {
                System.out.println(t + " " + phi);
            }
        }
    }
}

6.4 Matlab/Simulink — Cam Signal Skeleton

% Piecewise cam profile in MATLAB (usable as a Simulink "MATLAB Function" block)
function x = cam_profile(theta, theta_r, theta_h)
theta = mod(theta, 2*pi);
if theta < theta_r
    a2 = 1/theta_r^2;
    x = a2*theta^2;
elseif theta < theta_h
    x = 1;
else
    b2 = 1/(2*pi-theta_h)^2;
    x = b2*(2*pi-theta)^2;
end
end

% In Simulink:
% 1) Use a Ramp/Clock -> Gain (omega) -> cam_profile(theta)
% 2) Connect output to a Scope to visualize x(t).

These short codes are foundations for later lessons where we formalize actuation and sensing; for now, they reinforce that early automata were mathematically describable machines.

Problem 1 (Gear-train ratio): A three-gear train has tooth counts \(N_1=30\), \(N_2=75\), \(N_3=20\). Gear 1 drives Gear 2, and Gear 2 drives Gear 3. Find \( \omega_3/\omega_1 \) and the sign of the rotation.

Solution:

\[ \frac{\omega_3}{\omega_1} = (-1)^{2}\frac{N_1}{N_2}\frac{N_2}{N_3} = \frac{30}{75}\cdot\frac{75}{20} = \frac{30}{20} = 1.5. \]

There are two meshes, so the final sign is positive: Gear 3 rotates in the same direction as Gear 1, with 1.5 times speed.

Problem 2 (Cam smoothness): For the cam rise segment \(x=a_0+a_1\theta+a_2\theta^2\) on \(0\le\theta<\theta_r\), and the hold segment \(x=1\) on \(\theta_r\le\theta<\theta_h\), choose \(a_0,a_1,a_2\) such that \(x(0)=0\), \(x(\theta_r)=1\), and \(x'(\theta_r)=0\).

Solution:

Conditions: \(x(0)=a_0=0\). Next,

\[ x(\theta_r)=a_1\theta_r+a_2\theta_r^2=1, \qquad x'(\theta)=a_1+2a_2\theta. \]

Enforce \(x'(\theta_r)=0\Rightarrow a_1+2a_2\theta_r=0\Rightarrow a_1=-2a_2\theta_r\). Substitute into the position condition:

\[ (-2a_2\theta_r)\theta_r+a_2\theta_r^2=1 \Rightarrow -a_2\theta_r^2=1 \Rightarrow a_2=-\frac{1}{\theta_r^2}, \quad a_1=\frac{2}{\theta_r}. \]

Problem 3 (Escapement energy balance): Consider the driven oscillator \( I\ddot{\phi}+c\dot{\phi}+k\phi = \sum_{n}J\delta(t-nT)\). Assume a steady periodic regime with oscillation angular frequency \( \omega_{osc}\approx\sqrt{k/I} \). Show that the average injected energy per cycle equals the average dissipated energy per cycle.

Solution:

Over one period, the impulse at \(t=nT\) changes angular velocity by \( \Delta\dot{\phi} = J/I \), so injected kinetic energy is

\[ \Delta E_{in} = \tfrac{1}{2}I\left[(\dot{\phi}+\Delta\dot{\phi})^2-\dot{\phi}^2\right] = I\dot{\phi}\Delta\dot{\phi}+\tfrac{1}{2}I(\Delta\dot{\phi})^2. \]

In steady state the oscillation does not grow, so net energy change per cycle is zero:

\[ \Delta E_{in} - \Delta E_{diss} = 0 \Rightarrow \Delta E_{in} = \Delta E_{diss}. \]

Dissipated energy per cycle is \( \Delta E_{diss}=\int_{0}^{T} c\dot{\phi}^2\,dt \), demonstrating the balance.

Problem 4 (Rational approximations in clocks): Suppose a clock designer wants a speed ratio of exactly \(5/12\). Show how a two-stage gear train can realize this ratio using integer tooth counts.

Solution:

We need \( (N_1/N_2)(N_3/N_4)=5/12 \). Choose a factorization \(5/12=(5/6)(1/2)\). One valid set is \(N_1=50, N_2=60\) giving \(5/6\), and \(N_3=20, N_4=40\) giving \(1/2\). Thus the two-stage product gives the desired ratio. Many other integer choices exist by scaling tooth counts.

Early automata introduced three pillars of robotics in mechanical form: energy-driven actuation, geometry-based motion programming (cams and drums), and timing regularization (escapements). By modeling gears and cams mathematically, we see that “robotic thinking” existed centuries before electronics. The next lesson will move from preindustrial automata to the industrial revolution where these ideas scale into factory robotics.