Orthogonal Frequency Division Multiplexing (OFDM)
- Introduced by
- Weinstein and Ebert
- Year
- 1971
- Type
- multicarrier modulation scheme
- Family
- Signal processing
What does it do?
OFDM is a multicarrier modulation technique that divides a wideband channel into many narrowband orthogonal subcarriers.
Introduced by Weinstein and Ebert in 1971, it exploits the duality between time and frequency domains to efficiently use spectrum while mitigating intersymbol interference in frequency-selective channels. OFDM is now the standard for high-speed wireless systems including WiFi, cellular LTE, and digital broadcasting.
The intuition behind the method
Instead of sending all data on a single carrier (which suffers in fading channels), OFDM splits the signal into many parallel, closely-spaced carriers that are orthogonal to each other. Each carrier occupies only a fraction of the available bandwidth but together they use it efficiently. The key insight is that at the receiver, the orthogonality allows each carrier to be independently demodulated without interference from neighbors. This elegant use of the Fast Fourier Transform (FFT) makes OFDM computationally efficient and robust to multipath propagation.
History
First proposed by Weinstein and Ebert in 1971 but remained theoretical until the 1980s when FFT-based implementations became practical. The technique gained widespread adoption in the 1990s for digital broadcasting (DVB-T, DAB) and has since become standard in WiFi (IEEE 802.11a/g/n), cellular (LTE, 5G), and power line communications. Modern variants include Filtered-OFDM and Universal Filtered Multi-Carrier (UFMC) to reduce out-of-band emissions.
When to use it
OFDM is ideal for wideband systems with frequency-selective fading, especially when the channel has significant delay spread. It requires accurate synchronization and is sensitive to timing and frequency offsets. Prefer OFDM when spectral efficiency and robustness to multipath are critical; avoid it if strict latency constraints demand low processing overhead or if the channel is extremely sparse (where single-carrier techniques might be simpler).
Limitations and considerations
- High peak-to-average power ratio (PAPR) requires power-efficient amplification
- Sensitive to synchronization errors and frequency offsets
- Overhead from cyclic prefix reduces efficiency in short-duration bursts
- Requires explicit channel state information for equalization
Source conditions do not replace independent scientific review or validation by an analysis engine.
Strengths
- High spectral efficiency by using orthogonal subcarriers
- Robustness to multipath fading via cyclic prefix
- Low-complexity FFT-based implementation
- Flexible subcarrier allocation and adaptive loading per carrier
- Backwards compatible with existing channel models
Applications
- WiFi (IEEE 802.11a/g/n/ax) achieving multi-Gbps rates over home networks with multipath reflections
- Cellular systems (LTE/5G) providing reliable coverage in urban environments with non-line-of-sight propagation
- Digital television broadcasting (DVB-T) distributing broadcast content over shared spectrum to fixed and mobile receivers
- Power line communication (PLC) overcoming noise and attenuation in electrical grid channels
- Digital audio broadcasting (DAB) enabling regional radio service with single transmitter (Single Frequency Network)
How it works
How it works
Input data stream is segmented into blocks of symbols, with each symbol destined for one of the N subcarriers.
Serial-to-parallel conversion groups symbols; mapping (e.g., QPSK, QAM) applies to each subcarrier to produce frequency-domain samples.
Inverse Fast Fourier Transform (IFFT) converts frequency-domain symbols to time-domain samples; a cyclic prefix is inserted to eliminate intersymbol interference.
The OFDM signal is transmitted through a multipath fading channel; at the receiver, the cyclic prefix is removed and FFT is applied to recover frequency-domain symbols.
Equalization (zero-forcing or MMSE) is applied per-carrier; demodulation and parallel-to-serial conversion recover the original data stream.
Common pitfalls
- Underestimating synchronization challenges; even small timing misalignment causes subcarrier orthogonality loss
- Ignoring PAPR; insufficient amplifier linearity causes inter-carrier interference and performance degradation
- Assuming cyclic prefix length without validating channel delay spread; excess delay causes inter-symbol interference
- Not adapting subcarrier allocation; static loading wastes resources on deep fades
Frequently asked questions
Why is the cyclic prefix necessary?
The cyclic prefix ensures that each OFDM symbol period encompasses the entire channel impulse response. This makes the circular convolution property of the DFT applicable, allowing simple one-tap equalization per subcarrier. Without it, linear convolution would spread a symbol across multiple OFDM symbol periods, causing inter-symbol interference.
How does OFDM handle frequency-selective fading?
By dividing the wideband channel into narrowband subcarriers, each subcarrier experiences approximately flat (frequency-flat) fading. This replaces a complex frequency-selective equalization problem with simpler per-subcarrier gain correction, dramatically reducing receiver complexity.
What is the role of synchronization in OFDM?
Synchronization errors (timing or frequency offset) destroy the orthogonality between subcarriers, causing inter-carrier interference. Timing offset shifts the FFT window and loses the cyclic prefix benefit; frequency offset rotates subcarrier phase differently and leaks energy across carriers. Both must be estimated and corrected before symbol detection.
How can PAPR be reduced in OFDM?
Techniques include clipping (crude but simple), companding, tone reservation, tone injection, and selected mapping (SLM). Each trades off computational cost, PAPR reduction, and data overhead. Most practical systems use a combination; 5G allows filter bank multi-carrier variants that inherently reduce PAPR.
Related methods
Sources
Weinstein, S. B., & Ebert, P. M. (1971). Data transmission by frequency-division multiplexing using the discrete Fourier transform. IEEE Transactions on Communication Technology, 19(5), 628-634.
Open DOI referenceCopied.Alves, H., Nouri, M., & Latva-aho, M. (2015). Performance analysis of orthogonal frequency division multiplexing for wireless networks. IEEE Access, 3, 1627-1640.
Open publication referenceCopied.