📚 Eksplorasi & Komparasi Topik Dokumen
Pencarian topik
Hasil Dokumen
Ditemukan: 2 dokumenPARADOKS ADOPSI AI AGENTIK: HAMBATAN TERHADAP PENERAPAN PERUSAHAAN UNTUK ADOPSI YANG BERTANGGUNG JAWAB DI INDUSTRI PERANGKAT LUNAK
Munculnya AI Agentik menjanjikan akan adanya kelepasan potensi penuh yang tadinya diharapkan perusahaan dapat realisasikan dengan Generatif AI. Namun, adopsinya masih terbatas dengan tingkat keberhasilan penerapan produksi dan realisasi nilai yang rendah meskipun mendapatkan dukungan eksekutif yang kuat dan investasi yang substansial. Studi ini mengeksplorasi mengapa kesenjangan ini ada di dalam industri perangkat lunak Indonesia. Kami menggabungkan wawancara dari 16 peserta, fokus diskusi kelompok, dan survei terhadap 100 praktisi industri. Hasilnya menunjukkan bahwa meskipun peserta umumnya percaya AI Agentik bermanfaat, mereka memiliki kepercayaan yang sangat rendah dalam mendelegasikan tugas atau membiarkan AI Agentik beroperasi secara otonom tanpa pengawasan manusia. Empat hambatan utama diidentifikasi: Kesenjangan Nilai-Skalabilitas, Perangkap Otoritas, Kesenjangan Pengembalian Investasi (ROI), dan Spiral Kapabilitas-Risiko. Untuk mengatasi tantangan ini, studi ini mengusulkan AegisOps AI, platform tata kelola perusahaan yang dirancang untuk membantu organisasi menggunakan Agentic AI dengan aman dan efektif. Platform ini berfokus pada memastikan bahwa sistem AI tersebut mampu, dapat dipercaya, dan dapat dipertahankan, untuk membantu organisasi untuk menutup paradoks adopsi AI Agentik.
DEVELOPMENT OF AUTOMATION LABVIEW PROGRAM FOR ACCELEROMETER CALIBRATION
Laboratorium Dinamika LPIT ITB has been performing accelerometer calibration for industry to check the accuracy of sensitivity and frequency range of the sensor. This calibration consists of linearity and frequency response test. Currently, these accelerometer calibration is performed manually by operators, which consumed a large amount of time. Recognizing these inefficiencies of manual accelerometer calibration, this bachelor thesis aims to develop an automated accelerometer calibration program that significantly reduce the consumed time required for calibration. This automation program was developed by first doing manual calibration to identify which parts needed to be automated. The manual calibration requires the operator to control the generated signal from signal generator, read the signal amplitude, record the data, and process the data. These processes consumed a large amount of time. Generally, the total time required to perform the calibration is 1.5 to 2 hours depending on the operator skills. After understanding the manual calibration, the next step is to select the appropriate NI (National Instrument) for the calibration needs. Afterward, an automated calibration program is designed using LabVIEW software. Based on the identification from the manual calibration, the program is designed to be able to generate and acquire signal, record the data, and process the data. Additionaly, this program is designed to present the calibration results in the form of a certificate. To validate the accuracy of the automated program, manual calibration is performed again to be serve as a reference. Based on the testing results and the comparison between manual and automated calibration, it can be concluded that the accelerometer calibration automation program has been successfully developed. The sensitivity result of the calibrated accelerometer obtained from automated calibration has an error of 0.04% relative to the manual calibration. As for the frequency response test, it produce the same result both from manual and automated calibration. In contrast to the manual calibration, automated calibration only consumed 30 seconds for each acceleration amplitude and 30 seconds for each frequency.