---
title: 'PEH: A Diverse Acronym in Research'
url: https://www.emergentmind.com/topics/peh
type: topic
---

# PEH: A Diverse Acronym in Research

PEH is a context-dependent acronym used across several technically unrelated research literatures. In recent arXiv usage, it most extensively denotes **piezoelectric energy harvester** in electromechanical and structural-health-monitoring research, but it also denotes **people experiencing homelessness** in AI and social-policy studies, **Per-Edge Hypothesis** in kinodynamic motion planning, **piecewise exponential hazard** in Bayesian survival analysis, **pure energy harvesting** in SWIPT relaying, and **photoelectric heating** in galaxy-evolution modeling [2503.07462][2604.11703][2509.26339][1806.07048][2002.02859][2009.08078].

## 1. Lexical scope and disciplinary usage

The acronym is not stable across fields; its meaning is set almost entirely by local technical context. Electromechanical models, base excitation, capacitance, and harvested voltage indicate **piezoelectric energy harvester**. Knowledge graphs, shelters, PHC access, and stigma analysis indicate **people experiencing homelessness**. Multi-hypothesis search over changing cost maps indicates **Per-Edge Hypothesis**. State-space survival models with intervalwise hazards indicate **piecewise exponential hazard**. SWIPT relaying distinguishes **pure energy harvesting** from FD-SWIPT operation. ISM thermodynamics and FUV-driven gas heating indicate **photoelectric heating** [2503.07462][2604.11703][2509.26339][1806.07048][2002.02859][2009.08078].

| Expansion of PEH | Research area | Representative arXiv id |
|---|---|---|
| Piezoelectric energy harvester | Energy harvesting, SHM, sensing | [2503.07462] |
| People experiencing homelessness | AI for service access, health inequity, bias detection | [2604.11703] |
| Per-Edge Hypothesis | Kinodynamic motion planning | [2509.26339] |
| Piecewise exponential hazard | Dynamic survival analysis | [1806.07048] |
| Pure energy harvesting | SWIPT full-duplex relaying | [2002.02859] |
| Photoelectric heating | Galaxy evolution and ISM physics | [2009.08078] |

This distribution suggests that PEH functions less as a single concept than as a recurrent acronym family whose interpretation is domain-specific.

## 2. PEH as piezoelectric energy harvester

In the engineering literature represented here, PEH denotes a device that converts mechanical strain or vibration into electrical energy via the piezoelectric effect. The canonical forms are cantilever beams or cantilever plates with bonded piezoelectric layers, sometimes in bimorph configuration, sometimes as flexible diaphragms, and sometimes as bridge-mounted plate harvesters. The electrical observable is typically the load voltage \(v(t)\), with instantaneous power \(p(t)=v^2(t)/R_L\) and harvested energy \(E=\int v^2(t)/R_L\,dt\) over a window [2503.07462][2205.06949][2210.10540].

A recurrent modeling pattern is the coupled electromechanical system. In a base-excited SDOF cantilever model for bearing monitoring, the mechanical and electrical dynamics are written as
\[
m \ddot{x}(t) + c \dot{x}(t) + k x(t) + \theta v(t) = - m \ddot{y}(t),
\]
\[
C_p \dot{v}(t) + \frac{v(t)}{R_L} - \theta \dot{x}(t) = 0,
\]
so that resonance tuning and electrical loading jointly determine the band of vibration energy converted into voltage and accumulated energy [2503.07462]. In bridge applications, the same acronym denotes a more elaborate Kirchhoff–Love plate model discretized by IsoGeometric Analysis and reduced by Modal Order Reduction before Runge–Kutta integration, with Particle Swarm Optimisation used to maximize harvested energy under measured bridge accelerations [2205.06949]. Extending that framework to a real cable-stayed bridge showed that the position of maximum displacement in the relevant vibration mode corresponds to the best location for energy harvesting, and that traffic intensity shifts the optimal PEH fundamental frequency within a specific range of frequencies [2303.03620].

The term also increasingly denotes a **bi-functional** element rather than a harvester only. In bridge monitoring, a single PEH can act as both harvester and sensor in a Simultaneous Energy Harvesting and Sensing system, with voltage CWT images passed to AlexNet for traffic-speed classification; the study reports a trade-off between energy harvesting efficiency and sensing accuracy, with no single geometry optimal for both objectives [2205.06974]. In bearing-condition monitoring, PEHs tuned to defect-sensitive bands are used as embedded analog feature extractors, so the harvested energy itself becomes the diagnostic feature and a low-dimensional feature vector \(\mathbf{E}=[E_1,\dots,E_N]^\top\) replaces high-rate acceleration streaming [2503.07462]. In wearables, PEH-based insoles support simultaneous energy harvesting and gait recognition; the reported architecture achieved 12% higher recall, harvested up to 127% more energy, and consumed 38% less power than the stated state of the art [2009.02752].

Another major branch of PEH research concerns interface electronics and bandwidth extension. A shock-optimized SECE interface was reported with 30 nA quiescent current, 80 nW to 14 mW operating range, and 420% harvested energy improvement relative to a standalone full bridge rectifier under shocks [1803.07930]. A nonlinear simulated framework combined a Bouc–Wen PEH model, SECE, hybrid MPPT, and a switched-capacitor array; under variable excitation and load changes it reported 93–95% MPPT efficiency and up to 63% increase in harvested power when the excitation was far from resonance [2507.12163]. For broadband operation, BF electronics were paired with high \(\kappa_e^2\), exploiting the split between \(f_{sc}\) and \(f_{oc}\); for a measured device with \(\kappa_e^2=0.069\), \(f_{sc}=673.00\) Hz, and \(f_{oc}=696.00\) Hz, the reported 3-dB bandwidth with BF and a diode bridge was 50 Hz, or 7.4% of \(f_{sc}\) [1910.00557]. Flexible polymer implementations also fall under the same acronym: a PVDF-TrFe/PET circular diaphragm harvester reported a first resonance at \(9.74\,\mathrm{kHz}\), \(\zeta=0.117\), \(Q=4.284\), and \(V_{\max}=984\,\mathrm{mV}\) in a wind setup [2210.10540].

## 3. PEH as people experiencing homelessness

In AI, policy, and computational social-science papers, PEH denotes **people experiencing homelessness**, explicitly as a person-first term. The associated research focus is not a physical device but a structurally vulnerable population whose access to shelter, food, mental-health services, libraries, social security offices, and primary healthcare is mediated by information barriers, administrative exclusion, spatial access, and public stigma [2604.11703][2507.23644][2508.13187].

In service-navigation research, DreamKG is a knowledge-graph-augmented conversational system designed for PEH and helpers in Philadelphia. It grounds responses in a Neo4j knowledge graph, supports location-aware and time-sensitive queries, performs geocoding, radius-based search, temporal filtering over day-of-week and opening-hour nodes, and returns service cards plus map outputs. Preliminary evaluation reports **59% superiority over Google Search AI on relevant queries** and **84% rejection of irrelevant queries**, with the latter used as a safety indicator against hallucinated service recommendations [2604.11703].

In health-inequity modeling, PEH are the focal agents in a Capability Approach-guided agent-based and reinforcement-learning simulation for Barcelona. The proof-of-concept formalizes resources, conversion factors, capabilities, and functionings in an MDP where access to PHC depends on registration status; non-registered residents must engage with social services before PHC becomes feasible [2507.23644]. The capability metric is defined as
\[
\text{Central Capability}_{i}(t) =
\frac{ \sum_{(a_k, c)\in \mathcal{R}_{\mathrm{act}} \alpha_k \cdot a_{ik}(t) }{ \sum_{(a_k, c)\in \mathcal{R}_{\mathrm{act}} |\alpha_k| }.
\]
The reported result is that both registered and non-registered agents eventually learn policies reaching healthy terminal states, but the non-registered agent incurs more negative rewards, more actions, and a longer route through registration before receiving PHC [2507.23644].

In bias-detection research, PEH denotes the population targeted by homelessness stigma analysis. A manually annotated multi-modal dataset was compiled from Reddit, X, news articles, and city council meeting minutes across 10 U.S. cities, with a 16-category multi-label schema that includes harmful generalization, deserving/undeserving, NIMBY, government critique, societal critique, solutions/interventions, and racist content [2508.13187]. The gold standard contains 1,702 entities, average agreement is about 78.38% per category, and LLM evaluation shows that in-context classification by local models approaches closed-source systems while LLMs outperform BERT when averaging across all categories [2508.13187]. Across these studies, PEH is therefore a population descriptor tied to service access, capability deprivation, and discursive bias rather than a biomedical or engineering object.

## 4. PEH as Per-Edge Hypothesis

In kinodynamic motion planning, PEH stands for **Per-Edge Hypothesis**, the first multi-hypothesis method proposed for planning across inconsistent world models [2509.26339]. The setting is a mobile ground robot whose cost maps change across planning cycles, so a region may alternate between obstacle and free space. PEH extends KEASL and ARA\(^*\) by evaluating each edge expansion across a history of world hypotheses and invoking a rerouting sub-search whenever an edge is valid in some hypotheses but invalid in others [2509.26339].

The contrast with VEH is explicit. VEH accepts an expansion only if
\[
\forall i \in \{1,\dots,H\},\; \text{Trajectory}(n \to n') \text{ is collision-free in map } M_i,
\]
whereas PEH relaxes this to
\[
\exists i \in \{1,\dots,H\}\;\text{s.t.}\; \text{Trajectory}(n \to n') \text{ is collision-free in } M_i,
\]
and then uses rerouting in each invalid hypothesis to incorporate the cost of a hypothetical detour into the edge cost [2509.26339]. This makes PEH less conservative than VEH and more risk-aware than SH, because the planner prices in the potential cost of needing to reroute around previously hazardous regions.

The method’s limitation is computational rather than conceptual. Because PEH invokes a sub-search for every node expansion that crosses a divergence point, its runtime scales poorly with the number of divergent edges and hypotheses. The reported preliminary result is that **PEH and GEH are unable to generate solutions in less than one second**, which exceeds the stated requirement for field deployment; the field experiments therefore focus on SH, VEH, and GEGRH rather than PEH itself [2509.26339]. In this literature, PEH is thus an algorithmic design point in multi-model kinodynamic planning, not an energy harvester.

## 5. PEH as piecewise exponential hazard

In Bayesian survival analysis, PEH stands for **piecewise exponential hazard**. The model starts from a Cox-type multiplicative hazard
\[
\lambda(t \mid \mathbf{x}) = \lambda_0(t)\,\exp\{\mathbf{x}^\top \boldsymbol{\beta}(t)\},
\]
then partitions time into intervals \(I_j=[\tau_{j-1},\tau_j)\) and assumes both the baseline hazard and the regression coefficients are piecewise constant over those intervals [1806.07048]. The interval-specific hazard becomes
\[
\lambda_{ij} = \lambda_{0j}\exp(\mathbf{x}_i^\top \boldsymbol{\beta}_j),
\]
or, after reparametrization,
\[
\ln \lambda_{ij} = \mathbf{z}_i^\top \boldsymbol{\beta}_j.
\]
This produces a tractable likelihood that factors across intervals while allowing time-varying covariate effects [1806.07048].

The dynamic version embeds the sequence \(\boldsymbol{\beta}_1,\dots,\boldsymbol{\beta}_J\) in a state-space model with random-walk evolution,
\[
\boldsymbol{\beta}_j = \boldsymbol{\beta}_{j-1} + \boldsymbol{\epsilon}_j,\qquad \boldsymbol{\epsilon}_j \sim N(\mathbf{0}, \mathbf{U}_j),
\]
and performs Bayesian inference by particle smoothing rather than conventional MCMC [1806.07048]. The proposed algorithm uses three particle filters and constructs efficient proposal distributions through a Laplace approximation tailored to the posterior of the interval-specific hazard and linear Bayes updates for \(\boldsymbol{\beta}_j\). The reported outcome is an effective sample size **more than two orders of magnitude larger than a state-of-the-art MCMC sampler for the same computing time**, while scaling well in high-dimensional and relatively large data [1806.07048]. Here PEH denotes a survival-model family and is unrelated to the electromechanical usage.

## 6. Other specialized expansions

Two additional specialized expansions appear in the supplied literature. In SWIPT relaying, PEH denotes **pure energy harvesting**, one of the two operating modes in the harvest-and-opportunistically-relay protocol. The relay switches between PEH mode and FD-SWIPT mode depending on the direct-link SNR and its residual energy. In PEH mode the relay only harvests RF energy, with harvested energy
\[
E_{\text{PEH}} = \eta P_{\text{S}} |h_{\text{SR}}|^2,
\]
and it cannot transmit covert messages because there is no legitimate forwarding cover [2002.02859]. This usage is mode-specific and belongs to communication-theoretic energy management rather than piezoelectric harvesting.

In luminous-disk galaxy simulations, PEH denotes **photoelectric heating**, the heating of interstellar gas by electrons ejected from dust grains by FUV photons. The adopted volumetric heating law is
\[
n\,\Gamma_{\rm pe} = \beta\,F_e\,n\,G_0 \quad {\rm erg\,cm^{-3}\,s^{-1}},
\]
with \(F_e \in [0.003,\,0.05]\) as the photoelectric heating efficiency parameter [2009.08078]. In the reported Milky Way-like simulations, switching PEH on suppresses star formation from negligible values to approximately a factor of five, raises gas outflow rates and loading factors, and in gas-rich models suppresses disk-clump formation and bulge growth via clump migration [2009.08078]. This is again a complete semantic departure from the engineering and social-policy uses.

Across these literatures, PEH is therefore best understood as a high-collision acronym. Its interpretation depends not on the token itself but on the local formalism: electromechanical state equations, service-access ontologies, multi-hypothesis search, intervalwise hazard models, relay-mode control, or ISM heating laws.

Source: https://www.emergentmind.com/topics/peh