TL;DR
This paper proposes a new "maximal selection" (MS) spatial diversity scheme for 60 GHz millimeter-wave WLANs to combat human-induced shadowing. MS outperforms the existing "equal-gain" (EG) scheme by intelligently selecting the strongest available path, leading to higher link margins and reduced computational complexity, especially in frequency-selective channels.
What Problem Does This Paper Address?
The IEEE 802.11ad standard promises multi-gigabit per second (Gbps) wireless communication using the 60 GHz band. However, this high frequency band suffers from significant path loss and is highly susceptible to human-induced shadowing, which can easily disrupt the connection. Existing solutions like re-beamforming or multihop schemes are reactive and only detect a lost link after substantial data has been dropped due to the high data rates. A previous proposal, the "equal-gain" (EG) spatial diversity scheme, aimed to mitigate this by simultaneously forming multiple beams along the strongest paths. This paper revisits EG under more realistic frequency-selective (FS) MIMO channels and proposes a novel, suboptimal "maximal selection" (MS) scheme that offers improved performance.
Methods
The paper revisits the Equal-Gain (EG) diversity scheme, first deriving explicit expressions for its total power gain under a frequency-selective (FS) MIMO channel model. It then proposes a suboptimal spatial diversity scheme called Maximal Selection (MS). The core idea of MS is to actively "trace" the shadowing process. Instead of maintaining equal gains across all selected paths like EG, MS intelligently selects the path with the currently strongest channel gain. This selection is made possible by the high data rates of 60 GHz WLANs, which allow for rapid channel estimation and beam switching within the duration of a single packet. The MS scheme leverages a "shadowing tracing algorithm" to continuously monitor the primary path and quickly switch to an alternative path if the primary one becomes blocked.
Datasets
Not discussed in the paper.
Key Insights and Ideas
- Frequency-Selective Channels Matter: The paper highlights that previous work on EG diversity often assumed frequency-flat channels. For practical 60 GHz systems with their large bandwidth, frequency-selective channels are the norm, and EG's performance can be significantly impacted.
- Maximal Selection for Proactive Diversity: MS offers a proactive approach to shadowing. By continuously monitoring and selecting the strongest path, it avoids the reactive nature of other methods and minimizes data loss.
- Leveraging High Data Rates: The effectiveness of MS hinges on the multi-Gbps speeds of 60 GHz WLANs. This allows for rapid channel estimation and beam switching within the time it takes to transmit a single packet, making the shadowing tracing feasible.
- Reduced Computational Complexity: MS simplifies calculations compared to EG, particularly when dealing with multiple antennas and paths, by avoiding matrix inversions and multiplications.
- Mitigating Frequency-Selective Effects: By always directing beams towards the strongest path, MS inherently reduces the negative impact of frequency-selective fading.
Key Results
- Higher Link Margin: MS consistently achieves a higher link margin compared to EG, both in normal (non-blocked) and blocked scenarios. This improvement is more pronounced when the line-of-sight (LOS) path is not blocked.
- Improved Received Power: In blocked cases, MS provides significantly higher received power than EG. In non-blocked cases, MS also shows an advantage in received power, especially with larger reflection losses.
- Reduced Frequency-Selective (FS) Effect: MS exhibits less FS effect, leading to better Bit Error Rate (BER) performance, particularly in non-blocked scenarios.
- Computational Efficiency: MS offers lower computational complexity than EG due to simpler weight vector calculations.
- Minimal Efficiency Degradation: The proposed shadowing tracing algorithm incurs only a small efficiency degradation (around 1%) due to periodic pauses and beam switching, which is deemed acceptable given the large shadowing durations.
Limitations and Future Work
- Shadowing Tracing Cost: While deemed acceptable, the shadowing tracing in NLOS communication requires periodic pauses and beam switching, which can degrade communication efficiency. The paper suggests that selecting a relatively large estimation period can mitigate this.
- Suboptimal Nature: The paper explicitly calls MS a "suboptimal" scheme. While it outperforms EG, further research might explore truly optimal diversity schemes.
- Ideal Assumptions: The analysis in Section III makes ideal assumptions about non-overlapping reflection directions. While stated as reasonable for large antenna arrays, real-world scenarios might deviate.
- Future Work: The paper implicitly suggests further exploration of optimal gain settings and more sophisticated shadowing prediction or mitigation techniques. The performance evaluation is based on simulations and comparisons with existing schemes, implying that real-world experimental validation would be a logical next step.
Summary updated