Smart lighting has become a universal smart product solution, with global revenues of up to US $\$ $ 5.9 billion by 2021. Six main factors drive the technology: light-emitting diode (LED) lighting, sensors, control, analytics, and intelligence. The Internet of things (IoT) concept with the end device, platform, and application layer plays an essential role in optimizing the advantages of LED lighting in the emergence of smart lighting. The ultimate aim of smart lighting research is to introduce low energy efficiency and high user comfort, where the latter is still in the infancy stage. This paper presents a systematic literature review (SLR) from a bird’s eye view covering full-length research topics on smart lighting, including issues, implementation targets, technological solutions, and prospects. In addition to that, this paper also provides a detailed and extensive overview of emerging machine learning techniques as a key solution to overcome complex problems in smart lighting. A comprehensive review of improving user comfort is also presented, such as the methodology and taxonomy of activity recognition as a promising solution and user comfort metrics, including light utilization ratio, unmet comfort ratio, light to comfort ratio, power reduction rate, flickering perception, Kruithof’s comfort curve, correlated color temperature, and relative mean square error. Finally, we discuss in-depth open issues and future challenges in increasing user comfort in smart lighting using activity recognition.

This study aims to analyze the correlation between electrical properties of soil and NPK concentration in the soil. The conductivity and capacitance of soil with various NPK concentrations were measured. The main contribution in this study is to provide recommendations on the use of the corresponding electrical parameters for determining NPK levels in the soil. So far, the measurement of the electrical method generally uses the conductivity value for all conditions, even in wet and dry soil conditions. In fact, wet and dry soil conditions affect the conductivity. Therefore, other electrical parameters are needed that can correct the results of NPK measurements in the soil. Conductivity in relation to capacitance can show better measurement validity. In this study, the linearity of the relation between conductivity or capacitance to the addition of NPK levels in the soil was investigated. Soil samples were prepared by varying the mass concentrations of the elements N, P, and K separately. The conductivity and capacitance values for each NPK variation were measured and mapped. The results showed that the conductivity in dry soil conditions was more linear, with average coefficient of determination of the three elements being 0.81, compared to capacitance with a coefficient of determination of 0.72. Whereas the addition of water to the soil sample resulted in better sensitivity and linearity in the capacitance compared to the conductivity, with each average coefficient of determination being 0.55 and 0.34. These results provide alternative solutions using capacitance as an indicator for determination of NPK content in wet soil. The use of capacitance parameters combined with conductivity in determining NPK content in soil is a novelty contribution to this research and can be applied to the measurement technology in agriculture.

Our previous research applied a classification-integrated moving average (CIMA). This intelligence method improves the performance of passive infrared (PIR) sensors in smart lighting to make control more comfortable for the user. However, intelligence, synonymous with the use of the cloud and large latency, contradicts the real-time nature demanded by smart lighting. This paper proposes edge smart lighting (EdgeSL) architecture, an edge-computing architecture for real-time improved CIMA smart lighting control. We developed novel intelligent control using permutation importance feature selection to improve the CIMA algorithm. Three model compression methods for the k-nearest neighbor (KNN) model are compared, including the novel knowledge distillation on KNN called DistilKNN. The KNN model is the basis of CIMA. The method allows the model to run in an edge-computing environment. The experiment was carried out by evaluating processing time in three different environments, namely edge, fog, and cloud architecture. Our test results show that DistilKNN has the best accuracy compared to other methods, including pruning and quantization, which is 0.93. After deploying the compressed model to the NodeMCU, edge computing has a lower average processing time than fog computing and cloud computing, namely 9.0, 60.1, and 207.6 ms, respectively. After going through the Shapiro-Wilk test, we learned that the three processing times are not normally distributed. So after testing with the Wilcoxon test, it is proved that the EdgeSL has the best performance, where the average processing time with the other architectures has significant differences.

One of the most challenging issues in IEEE 1609.4 is the assurance of Quality of Service (QoS), i.e. to improve throughput and reduce delay in the sublayer Medium Access Control (MAC) IEEE 1609.4. The prioritization of each service package, using Enhanced Distributed Channel Access (EDCA) at the MAC sublayer is designed based on the IEEE 802.11e with some modifications to the transmission parameters.

The Quality of Service (QoS) on the traffic environment with high mobility, dynamic topology, and change trajectory is one of the challenges in the use of Vehicular Ad-Hoc Networks (VANET) Routing Protocols. PUMA routing protocol presents multicast work mechanism based on shared mesh topology. It always dynamically adapt to VANET traffic conditions. In this paper, we evaluated the PUMA routing protocol performance using the Manhattan Mobility Model with the effect of the Nakagami fading distribution in VANET. We also analyse the average node distribution in the tagged vehicle communication range using Poisson distribution. We use the NS 2.34 and VanetMobiSim for the simulation and evaluation of the performance. The performance results will be analyzed by varying the traffic parameters in the Manhattan mobility model from low to high traffic condition. The simulation result shows that the communication range under the Nakagami fading distribution affected the Quality of Service (QoS) in VANET. We found that the delay is higher compared with those in the condition without Nakagami fading. The throughput is also lower compared with those in the condition without Nakagami fading.