4.18 Information Signal Error

An information signal error is any discrepancy between a transmitted signal and the corresponding received signal that causes the receiver to decode a message different from the one that was sent. In cybernetic communication theory, signal errors are fundamental obstacles to reliable information transmission and are analyzed mathematically through probability theory, coding theory, and channel models. Understanding the sources, types, and consequences of signal errors is essential for designing communication systems that meet reliability requirements.

Signal errors arise from the interaction between the transmitted signal and the imperfections of the channel through which it travels. The primary causes include thermal noise, which is the random motion of charge carriers in electronic components; interference from other signals sharing the same medium; fading due to multipath propagation in wireless channels; impulse noise from transient electrical disturbances; and channel distortion that alters the waveform in systematic ways. Each of these causes produces a different statistical profile of errors and requires different mitigation strategies.

The simplest model for analyzing signal errors is the binary symmetric channel, where each transmitted bit is independently flipped with probability p. For this channel, the conditional probability of receiving a 0 when a 1 was transmitted, and vice versa, is equal to p. The overall bit error probability is simply p, and the information capacity of the channel is:

where H(p) is the binary entropy function. As p increases from 0 to 0.5, the capacity decreases from 1 bit per channel use to 0, reflecting the increasing corruption of each transmitted symbol.

For analog signals in the presence of additive white Gaussian noise, the error is not a discrete flip but a continuous perturbation. The received signal y equals the transmitted signal x plus a Gaussian noise term n:

When the receiver must make a binary decision based on the received signal, the probability of error depends on the signal-to-noise ratio. For binary phase-shift keying (BPSK) with signal amplitude A and noise variance σ², the bit error probability is:

where Q is the Q-function, the tail probability of the standard normal distribution.

Error types differ in their statistical structure and in the strategies required to combat them. Bit errors are isolated, independently occurring errors affecting individual symbols. Burst errors are correlated, affecting consecutive bits in a block due to a transient interference event such as a lightning strike or a deep fade in a wireless channel. Codes designed to combat random bit errors, such as Hamming codes, may perform poorly against burst errors, whereas interleaving and burst-error-correcting codes, such as Reed-Solomon codes, are specifically designed for bursty error patterns.

Erasure errors occur when a receiver knows that a particular symbol has been lost or corrupted beyond recovery, rather than receiving a wrong symbol without knowing it. Erasures provide strictly more information than errors to the decoder: knowing which positions are unreliable allows the decoder to concentrate its error-correction resources on those positions. The capacity for erasure channels is correspondingly higher than for error channels with the same fraction of affected symbols.

Error detection codes identify when errors have occurred without necessarily correcting them. Cyclic redundancy checks (CRC) append a small block of redundant bits to each frame, computed as the remainder of a polynomial division over the message bits. Any error pattern that alters the received polynomial in a detectable way will cause the CRC check to fail, alerting the receiver that retransmission is needed. This approach is used extensively in networking protocols where retransmission is practical and the overhead of full error correction would be unnecessary.

Error correction codes, by contrast, add sufficient redundancy to allow the receiver to both detect and correct errors. The minimum Hamming distance of a code determines its error-correction capability: a code with minimum distance d can detect up to d−1 errors and correct up to ⌊(d−1)/2⌋ errors in any received codeword. Modern codes such as turbo codes and low-density parity-check (LDPC) codes operate near the Shannon limit, correcting errors with overhead that is only marginally more than the theoretical minimum required.

In the broader cybernetic framework, signal errors represent failures of the communication channel to faithfully transmit the state of the source to the destination. These failures propagate into the feedback loops of control systems, introducing noise and uncertainty into the controller's estimate of the plant's state. Cybernetic system design must account for the statistical properties of signal errors to ensure that control performance remains adequate under realistic channel conditions, whether those channels are electronic circuits, sensory pathways in organisms, or information flows in organizations.