Inductive loop detectors (ILDs) are one of the most widely deployed traffic sensors. At present, for lane-by-lane detection, ILDs require separate connecting cables for each loop (each lane) and separate data acquisition systems or detector channels to process them. This becomes problematic with limited conduit and cabinet space. In most cases, transportation agencies use ILDs connected in series to avoid these constraints, in which case the lane-by-lane information is lost. However, research has shown that lane-by-lane detection can lead to safer and more efficient operations at signalized intersections. In order to ease the application of lane-by-lane detection, the current study proposes a solution that uses electronic circuit modification to convert the existing serially connected loops to carry out lane-by-lane detection. This system achieved 100% accuracy of lane-by-lane detection in test runs. The paper also proposes an improved loop design, for future installations, that can be used for vehicle classification and wrong way detection. The study implemented machine-learning algorithms for vehicle classification and direction determination with an accuracy of 99.6% and 78.57%, respectively, using single loop configuration.